2023 Investor Day｜Tesla

29:20Good afternoon, everyone.

29:22Good afternoon.

29:26My name is Zach Kirkhorn.

29:28I lead finance here at Tesla and welcome to our investor day.

29:39For those of you here in the audience at our Global Headquarters

29:43in Austin, Texas, we welcome you.

29:45Thank you for being here and thank you for traveling in.

29:48For those joining us virtually, thank you

29:50for being a part of the day today.

29:52So as we reflect back on the history of the company,

29:56There's been distinct phases of product advancement,

30:00technology innovation, and rapid volume growth.

30:04The most recent of which has been the global expansion

30:07and localization of the Model Y program.

30:10And so today, we want to talk about the future.

30:14We don't want to talk about this quarter or next quarter.

30:17We want to go further out into the future.

30:20And we've divided today's presentation into three parts,

30:24The first of which, we're going to go macro.

30:27What does it take to convert Earth to sustainable energy

30:32generation and use?

30:34The second, we want to talk about Tesla's contribution

30:38to that global need.

30:40We're going to go function by function through the company

30:44and you'll meet our entire leadership team.

30:47And we're going to get into the details of what those teams

30:50are doing as part of the broader goal.

30:54And then in the third part, we're going to bring it back up and talk about what this all

30:59means for the company as a whole.

31:03So before we get started, statements made in this presentation are forward-looking statements

31:09that are subject to risks and uncertainties.

31:12More details can be found in our written materials.

31:15So with that, let's get started.

31:18Part 1, Elon Musk and Drew Baglino.

31:34Master Plan Part 3.

31:37So as Zach was mentioning, the thing that I think is we wanted to convey, probably more

31:46importantly than anything else that we talked about here, is that there is a clear path to

31:52a sustainable energy earth.

31:55It's not, it doesn't require destroying natural habitats, it doesn't require us to be austere

32:09and stop using electricity and sort of be in the cold or anything.

32:14The story, and I think if this holds together quite well

32:17and will be actually publishing detailed white paper

32:20with all of our assumptions and calculations,

32:22is that there is a clear path to a fully sustainable Earth

32:26with abundance.

32:27In fact, you could support a civilization much bigger

32:31than Earth, much more than the 8 billion humans

32:35could actually be supported sustainably on Earth.

32:40And I'm just often shocked and surprised

32:43by how few people realize this.

32:50Most of the smart people I know actually

32:53don't see this clear path.

32:55They think that there's not a path to a sustainable energy

33:01future, or at least there's not one

33:03that is sustainable at our current population,

33:08or that we'd have to resort to extreme measures.

33:12None of this is true.

33:14So we're going to walk through the calculations

33:17for how to create a sustainable energy civilization.

33:24And to set the stage, today, our energy economy,

33:28let's be honest, it's sturdy and it's wasteful.

33:31Over 80% of global energy, primary energy,

33:33comes from fossil fuels.

33:34and only one third of that global energy

33:37actually ends up delivering useful work or heat.

33:40This is the problem statement,

33:41but we're here to talk about the solution.

33:44Yeah, it's like, for some of this I'm gonna elaborate

33:48because there's a very wide range of technical expertise

33:52out there from people who are like, you know,

33:55whatever level nine wizards in this subject

33:58to people who do not do engineering at all.

34:01So if you have a gasoline car, you're converting less than a third, often maybe only 25% of

34:12the energy in the gasoline is converted into motion.

34:15The rest is turning to waste heat.

34:17That doesn't do any good at all.

34:19And there's a lot of energy required even to get the oil out of the ground to refine the

34:23oil and to transport the gasoline to the gas station.

34:27So when you look at all that, for a typical gasoline car is actually going to be using

34:34less than 20% fully considered of the energy from the oil actually goes into motion.

34:42So this is a, when we see people doing calculations for what does it take to create a sustainable

34:48energy earth, they assume that the same energy amount is required for an electrified civilization

34:56versus a combustion civilization.

34:59This is not true because most of the energy

35:03a combustion is waste heat.

35:06And even to get the fuel to combust in the first place

35:09and get it to the end use, there's a lot lost along the way.

35:12I mean, this is the primary energy consumption, 165

35:15petawatt hours a year.

35:16Petawatt hour is a trillion kilowatt hours.

35:19So it's a large amount of energy.

35:21But the nice thing about electrified economy,

35:25There's a better way, we're gonna talk about it,

35:27is that through end use efficiency

35:30and through efficiency along every step of the way,

35:32actually the total energy use, it halves.

35:36So this is one of the most enabling aspects

35:38of electrifying everything,

35:41is that the sustainable energy economy

35:43is that much easier to accomplish.

35:45It's actually half the problem statement

35:47of the fossil fuel economy.

35:48Yeah, and we're being conservative here,

35:49so it could be better than half,

35:51but we're trying to have assumptions

35:55that are reasonable, not overly optimistic,

35:56in fact slightly pessimistic.

35:59So it's really better than half,

36:00but just say for, it's easy to make the argument

36:03that we need half as much energy

36:05within the electric economy versus a combustion economy.

36:08Yep.

36:11So how the master plan works, you wanna talk about?

36:14Yeah.

36:15So the thing that is needed at very large scale

36:20needed at very large scale that is not currently present is a vast amount of battery energy

36:26storage. Our rough calculations are that this is about 240 terawatt hours or 240,000 gigawatt

36:35hours. This is a lot of batteries but it is actually a very achievable amount. We're going

36:43to details on that. So that's a combination of electric vehicles and stationary storage.

36:49So if you've got solar or wind, you've got to store the energy when the wind is not blowing,

36:53the sun is not shining.

36:56And so we're assuming sort of an 8 to 1 ratio of stored energy to power.

37:02So 30 terawatt hours of power.

37:0530 terawatts of power.

37:08Our actual capital expenditure calculation for manufacturing investment is more like 6

37:17trillion.

37:18we made it higher to make it 10 trillion.

37:22And this is across mining, refining,

37:25battery factories, recycling, vehicle factories,

37:28all the things that we're gonna talk about

37:30needing to invest in to build

37:31the sustainable energy economy.

37:33Yeah, now, if you look at the total world economy,

37:36it's just under 100 trillion.

37:39So if this was spread out say over 10 years,

37:42it would be 1% of the global economy.

37:45Over 20 years, it would be half a percent.

37:47very global economy.

37:48So this is not a big number relative to the global economy.

37:55As Drew mentioned, you need about half as much energy

37:58with an electric economy versus a combustion economy.

38:01And in terms of wind and solar, how much land would be used,

38:05it's less than 0.2% of the land area of Earth.

38:11Generally, people don't realize quite how much energy

38:14is reaching us from the sun.

38:17It's roughly a gigawatt per square kilometer.

38:21And it's not as in China all the time,

38:23but if you multiply that by, say,

38:284 to get the continuous power, 4 or 5,

38:33then that gives you the land area of solar.

38:36And you can put wind and solar often in the same place.

38:38So all the places that currently have wind,

38:40you could have solar there and you doubled your energy.

38:43You can also put wind off shore.

38:45It doesn't even need to be on land.

38:46So when does even more flexible.

38:48It puts all of them short too.

38:50Yep.

38:50So Earth is 70% water.

38:53Anyway, what is that with a pretty, really,

38:57a remarkably small amount of Earth's land area,

39:00we can go fully sustainable.

39:04Yeah.

39:05And from a, do the resources and raw materials

39:08exist to support this transition?

39:10We'll go through that in detail,

39:11but we do not see any insurmountable resource

39:14challenges at all.

39:14In the end, we should be mining less ore to accomplish this economy than we currently do

39:20with the fossil fuel economy.

39:21And we're going to talk through that.

39:22Yeah, just emphasize that again, the electrified economy will require less mining than the current

39:29economy does.

39:30Less.

39:31Not more.

39:32Okay.

39:33So this is the plan, and now we'll get into a little more of the details of the plan.

39:39Basically five areas of work.

39:42First, renewable power, the existing grid.

39:45Second, switch to electric vehicles.

39:48Third, switch homes businesses and industry heating

39:51to heat pumps.

39:53Fourth, high temp heat delivery and storage

39:57for high temp industrial and chemical processes.

40:00And a little bit of green hydrogen in there

40:02for chemical processes that need hydrogen.

40:05And finally, sustainably fuel planes and boats.

40:08These are the five areas.

40:09And we're going to go into detail on all of them.

40:11Yeah, I mean, my personal opinion is that as we improve the energy density of batteries,

40:16you'll see all transportation go fully electric with the exception of rockets.

40:23That's awkward.

40:25But you can make the fuel with CO2 and water.

40:30So you can make methane with CO2 and water.

40:33And you can do that with just electricity.

40:35Yes, exactly.

40:36So, in fact, on Mars, we hopefully get there at some point, the atmosphere is CO2 and there's

40:44water ice throughout Mars.

40:46So you can take the CO2 and H2O and turn that into CH4, which is methane and oxygen.

40:52So ultimately, even rockets can be electrified.

40:58So first, repowering the existing grid with renewables, and this is going to be a consistent

41:02theme.

41:03You'll see our estimates for the number of terawatt hours, terawatts, and trillions of investment

41:07at the bottom of the page.

41:09This is already actively occurring in front of us.

41:1360% of the generation added to the US grid

41:15was solar in 2022.

41:17And actually, on a year-on-year basis,

41:19solar deployment is growing 50% year-on-year as of 2022.

41:24So this is a serious upswing.

41:29And if we continue this trend, this

41:30is going to be behind us before we even know it.

41:34Second, switching to electric vehicles.

41:36Again, 21% reduction in fossil fuel use by doing this alone.

41:41Obviously, Tesla is heavily engaged in this activity,

41:44as along with many others.

41:46Overall, EV production grew 59% year on year in 2022.

41:50And EVs hit an amazing 10% market share.

41:53I mean, it's an awesome milestone.

41:55I'm super excited to see that.

41:57And I got to.

41:59Yeah.

42:00This is obviously happening very rapidly.

42:06And I think really all cars will go to fully electric

42:11and autonomous.

42:14And so riding a non-autonomous gasoline car

42:19is gonna be analogous to riding a horse

42:21and using a flip bone.

42:22That's basically gonna be the situation.

42:25And we actually took a somewhat conservative assumption here

42:27in terms of how many batteries are required

42:29because the more the fleet is autonomous,

42:31the fewer, the smaller the fleet needs to be,

42:35just from a utility-pray basis.

42:36So we're not accounting for all of those benefits,

42:38or really much of those benefits at all in this number.

42:41And what does this fleet look like?

42:43Just rough view from our perspective,

42:46of course, we could be wrong,

42:47but you can see the sort of breakdown of the fleet

42:51by millions of vehicles.

42:53Our goal is to do 20 million electric vehicles a year.

42:57Yeah.

42:58Fewer vehicles will be needed at least passenger vehicles

43:02with autonomy.

43:03So this debate as to what that number is,

43:08but it's a number less than the number of vehicles

43:11needed today.

43:11There's roughly 2 billion cars and trucks

43:15in operation in the world today.

43:17Yeah, so what we show here is actually, I think only 1.4

43:20million or so.

43:22So we're 1.4 billion, I mean, or so.

43:25So a smaller fleet.

43:27And the numbers here in this presentation

43:29are around 85 million vehicles a year

43:32produced, just to give you a sense of how we're thinking about this.

43:34Again, we're going to put all these assumptions up online and encourage people's thoughts.

43:39Yeah.

43:40So, yeah.

43:43So we're basically heading rapidly towards an electric or autonomous future.

43:48Exciting.

43:50And one of the reasons why EVs are so enabling is this end use efficiency point.

43:56Tesla Model 3, it's four times more efficient

43:58from the well to wheel than a Toyota Corolla.

44:02And that's all about the efficiency of getting

44:04the electricity into the car in a sustainable energy economy

44:08and then how efficient the car is

44:10in transferring that stored energy to motion on the road.

44:14When compared to the engine in the Toyota Corolla

44:17and all the extraction refining transmission distribution

44:20of the gasoline to the Toyota Corolla.

44:22And just for, this is a fun reference.

44:25Model 3 can drive over a mile on the energy it takes to boil a pot of water for pasta,

44:30and then it can drive another mile on the energy it took to cook the pasta.

44:35And that pasta is one pound, and all three is like 4,000 pounds.

44:39Just to give you a sense of just like, it really doesn't use a lot of energy

44:43to move a Model 3 at 4,000 pound object down the road.

44:47Also heat is a lot more energy than motion.

44:49Yeah.

44:50Yeah.

44:51But people, you just boil a pot of water, you don't even think about it.

44:55It's just interesting how efficient these cars are.

44:58Next, switching to heat pumps in homes, businesses,

45:01and industry.

45:03Right now heat pumps meet 10% of building heating needs,

45:06install rates growing 10% year on year.

45:08It really needs to accelerate.

45:10Heat pumps can serve heat applications up to 200C

45:13in businesses and industry.

45:15And from an investment perspective,

45:16as you can see on this page, is actually

45:18the lowest hanging fruit in terms of displacing fossil fuels.

45:22You might be saying, what exactly is a heat pump?

45:24So heat pumps don't move heat.

45:29Don't create heat.

45:31They move heat.

45:33When you think about the natural gas

45:35furnace in your house, it is generating heat itself.

45:37But what the heat pump is doing is actually

45:39moving heat from outside of your house into your house.

45:41They're just an air conditioner or a refrigerator in reverse.

45:45So we're surrounded by heat pumps.

45:46There's like, they're all over this factory.

45:49They're in your house.

45:50And all this really is is about bringing them

45:52to displace all the fossil fuel heating

45:55in all the homes business in the industry that we can.

45:58And from an end-use efficiency perspective,

46:00there's a three times reduction

46:02in the total energy required to heat these buildings.

46:06So a real obvious thing to do.

46:08Yeah, heat pumps.

46:09There are not cars now as default,

46:12and at some point we might make heat pump for home.

46:16Yeah, maybe, maybe.

46:18Maybe.

46:20Next, a little bit more detail on electrifying high temp

46:24sort of industrial chemical processes.

46:26So over 50% of industrial heat is greater than 400C.

46:30Cement, steel, fertilizers, chemicals, plastics, metals

46:33refining, all need like 1,500C.

46:35So we need a solution here.

46:38Ultimately, it's purpose-built equipment

46:40that enables electrification.

46:42Carbon, graphite is stable up to 2,800C.

46:45There's other options in the 1,500C range

46:48like silicon carbide, other materials.

46:50So the idea here is you create and store heat

46:54when renewable power is available.

46:55If this is a sustainable energy economy,

46:58renewable energy is intermittent.

47:00Peak of the day, you've got more generation than you need.

47:02You make a bunch of heat then, and then you transfer

47:04that heat into the industrial process 24 hours a day

47:08using the stored heat you created

47:09when the sun or the wind was blowing.

47:11That's the concept here.

47:12And then on the hydrogen side, we also need green hydrogen

47:15to decarbonize metals and chemical refining processes.

47:18This is things like ammonia, making steel.

47:22There's roughly 120 million tons of hydrogen source

47:25from fossil fuel today to do these things.

47:27And hydrogen can also directly replace coal,

47:30which is currently used in a ton of steel production,

47:33through a process called directly reducing iron.

47:35You can replace blast furnaces

47:37with a hydrogen reducing direct reduce iron furnace.

47:42And this is the way to illuminate fossil fuels

47:44from these aspects of the economy and the CO2 associated with them.

47:48Yeah, so I mean, some of this room to disagree,

47:52but some amount of hydrogen is needed for industrial processes.

47:57My personal opinion is that hydrogen will not

47:58be used meaningfully in transport,

48:02but it shouldn't be.

48:04If you're going to use a chemical fuel,

48:06you should use CH4, not H2.

48:10But it's not unless it is needed for industrial processes.

48:14and can be produced just by splitting water, essentially.

48:18I mean, some of it's been done for decades and decades.

48:20This is not rocket technology.

48:25And lastly, a small part of the pie,

48:28but a necessary part of the pie,

48:29is sustainably fueling planes and boats.

48:32Shipping accounts for 3% of Global CO2.

48:35It's ripe for electrification.

48:37Even with lithium iron phosphate,

48:38long haul ships can be fully battery powered.

48:41So that's a great opportunity to electrify energy density

48:46is a little bit harder for planes,

48:47but short haul is doable today with some improvements.

48:49We'll get long haul underway, but even in the meantime,

48:52we can leverage sustainable aviation fuels

48:55produced and stored using access or noble electricity.

48:57There's a lot of work going on in this space.

48:59And it's not, yeah.

49:01Yeah, I mean, to really get long range aircraft

49:06and long range shipping to use lithium ion,

49:12you need to redesign the ship and not just,

49:14or the plane, and the plane,

49:18to take advantage of the fact that it is a new

49:22source of energy.

49:24It's a different architecture.

49:26So, just like with an electric car,

49:29you wouldn't just take a gasoline car

49:32and stick a battery in it.

49:33That's very sub-optimal.

49:36It's much more efficient to have the battery be the structure of the car.

49:42And make it mass efficient and optimize for battery.

49:48If that's done with aircraft, you can get long-range aircraft

49:53with sales at around 450 watt-hours per kilogram, which you can buy it right now.

49:59Actually, it's expensive, but I think that price will come down.

50:02So when we stack up all of these efforts, we end up with the numbers we shared at the

50:08beginning of the presentation, 30 terawatts, 240 terawatt hours, $10 trillion.

50:14And you may be saying, like, I need some context, is this feasible?

50:18Spoiler alert, it's entirely feasible.

50:23Just looking at it from a growth rate perspective, how much do we need to grow the deployment

50:28of these technologies?

50:29We're talking about only a 3x growth rate in solar and wind deployment.

50:35Solar is already growing at a breakneck pace as is wind.

50:39This gap is going to be closed really quickly.

50:41When we look at the electric vehicles, they have to grow 11x.

50:45Well, they grew 60% year on year last year.

50:48That growth rate is also going to close pretty darn.

50:51That gap is going to close pretty quickly as well.

50:53And lastly, storage.

50:56You know, Tesla's energy storage business has grown at 65%

50:59CAGR since 2016.

51:02The global energy storage business is accelerating pace as well.

51:06I mean, all of these gaps are going to close,

51:08especially as this momentum of the transition

51:10to sustainable energy accelerates.

51:14And of course, our goal on this page is 20 million EVs

51:17per year and one terawatt hour of stationary storage per year,

51:21basically as soon as we can.

51:23And then, you know, what about this investment?

51:28How do I have a reference point on this investment?

51:30You know, Elon mentioned it's 10% of, you know,

51:34one year's world GDP.

51:37Another way of thinking about it is how does it compare

51:39to what we're investing, like what we invested last year

51:41in the fossil fuel infrastructure.

51:44And it's 60% of that investment.

51:46So actually building this sustainable energy economy is less

51:51than extending the fossil fuel economy from a year over year investment basis. So, very

51:57doable. When we look towards, does this fit on the planet? Absolutely. Less than 0.2% of

52:04land. As a reference point, the total land area intensively farm today is 12.5% of all

52:13land. So, I mean, you drive around, do you see some farms, but you don't see them everywhere.

52:18This is an order of magnitude, more than order of magnitude difference between farming

52:22and what we're talking about for sustainable energy land.

52:25And it doesn't need to displace farmland or forests or jungle or any ecological preserve.

52:35It can be used in very sparsely populated desert regions.

52:40Baran areas that are just not really fit for development or otherwise used.

52:45Yeah.

52:46I mean, 0.2% can fit into a lot of places.

52:48Yeah, it's essentially no meaningful ecological impact.

52:53In fact, transition to a sustainable energy economy

52:57would result in a substantial reduction

52:59in current ecological impact.

53:02It's a great way to put it.

53:03Yeah.

53:04And what about on the mineral extraction side?

53:07So this is a cartoon that sort of gives you a sense

53:09for all the ore and the extracted minerals

53:13that are coming out of the earth every year.

53:16It's about 68 gigatons.

53:20So each truck has a gigaton.

53:21What does this look like when we're

53:22in a sustainable energy economy?

53:25Looks like that.

53:27The fossil fuel extraction disappears.

53:29We replace it with the materials required

53:31to fulfill the sustainable energy economy.

53:34It actually reduces.

53:35Now, it's not to say that we don't need to continue

53:38to explore, bring our mining and refining

53:40for the specific materials for the sustainable energy economy.

53:45We do.

53:46But the investment in mass flows are all very achievable,

53:48just looking at what is already happening on the planet.

53:50Like this is nothing out of scale

53:52of what has been done and is already being done.

53:56And then we calculated on a sort of, you know,

53:59element by element basis,

54:01the resources are there to support the transition.

54:04You know, this is cumulative demand

54:06to move in the sustainable energy economy direction

54:08until 2050 relative to USGS resources today.

54:12You know, we're not breaking the resource bank

54:16for any of these materials.

54:17And then when we look at what really happens

54:19as we move forward, history teaches the more we look,

54:23the more we find.

54:24What people think happens is, oh, there's

54:26this many resources, next year there's

54:27going to be less, because we're going to extract them.

54:29What actually happens is, as we extract resources,

54:33we find more.

54:35And you can see, on the right, what

54:36has actually occurred with the key materials

54:38to the sustainable energy economy.

54:40Since 2000, as the sustainable energy economy has been growing,

54:43and Tesla's been growing, and all the industries around us

54:45have been growing, the actual resource availability

54:48has increased, not decreased.

54:50Yeah, this seems to be quite a bit of confusion

54:54about lithium.

54:55Lithium is extremely common.

54:57It's one of the most common elements on earth.

54:59There's no country that has monopoly on lithium,

55:03or even close to it.

55:07There's enough lithium, or in the United States,

55:10to electrify all of earth if the United States was the only

55:13place producing lithium. There's enough domestic material to electrify Earth. It's very common.

55:22The limiting factor is the refining of the lithium into battery grade lithium hydroxide

55:29or lithium carbonate. That's the actual limiting factor.

55:32And the same is true for these other materials. And this is, again, these are not like crazy

55:36technologies. It's just the investments need to be made and the investments, they're not

55:40gigantic they just need to happen.

55:42Right.

55:43Nichols maybe the, of them all the trickiest one to solve,

55:47but as we showed with the graph there,

55:50maybe you need like 30% of the world's

55:51known nickel reserves.

55:53And the nickel reserves have actually grown,

55:55since 2000.

55:56There is more, yeah, exactly.

55:57So, and you only need nickel for basically aircraft,

56:03long range boats, and very long range cars or trucks.

56:08but the vast majority of the heavy lifting

56:12for electrification will be iron-based cells.

56:17And the Earth is, iron is actually literally

56:19the most common elements on Earth.

56:21So it's a little trivia point.

56:25If you say, what is Earth made of?

56:27By mass, it is made of iron more than anything else.

56:31And second, oxygen, and then everything else after that.

56:36So basically where a muddy rust ball is, what Earth is.

56:43So an iron cathode is sort of, you're definitely

56:46not going to run out of iron.

56:46There's so much iron is insane.

56:54Yeah, and ultimately, this resource extraction,

56:58we go through this effort, we build these batteries,

57:02and then we recycle these batteries.

57:03So ultimately, we're doing this to build this

57:07sustainable energy economy, but the maintenance amount of

57:10ore that we require is really an order of magnitude

57:13or more less because of recycling.

57:17So in the end, a sustainable energy economy

57:21is within our reach and we should accelerate it.

57:24Yeah, I mean, so this is really the main message of today.

57:30And I really wanted today to be not just about

57:32tells the investors who own stock, but really anyone who is an investor in Earth.

57:38And what we're trying to convey is a message of hope and optimism.

57:45And hope, optimism that is based on actual physics and real calculations, it's not wishful

57:52thinking.

57:53Earth can and will move to a sustainable energy economy, and will do so in your lifetime.

58:01Thank you.

58:03Yes.

58:04Thank you.

58:05And I just want to welcome Lars and Franz to the stage.

58:24Hi, I'm Franz.

58:25I have lead design at Tesla.

58:28And I'm Lars.

58:29I've been doing vehicles with Franz for almost 13 years now.

58:34So I joined Tesla in 2008 to vertically integrate design into the company.

58:41It didn't exist before.

58:43It was a pretty small team.

58:44It was tasked with designing most beautiful, innovative, and well-engineered vehicles on

58:49the planet.

58:50No small task in 2008.

58:52There is not a lot going on on the EV sector then.

58:55Since then, we focused on constant improvement and cost, efficiency, innovation, things that

59:00you'll continue to hear about today, while continuing to design the most desirable cars.

59:08Today we produced cars differently than we did 10 years ago.

59:10But the end result is always an exciting, futuristic, and desirable set of vehicles.

59:16Back then, we only had a handful of designers and engineers like myself, but we had a great

59:19vision to radically change transportation.

59:24So back in 2008, where we were designing Model S, we didn't have a factory.

59:30In fact, we had a really small engineering team and a tiny design team.

59:36But that allowed design to lead all the conversations.

59:39It let us innovate forward thinking ideas like how do you fit seven people into a sedan?

59:46Or how do you make door handles disappear into the doors?

59:50Or putting a huge touchscreen into the center of the vehicle, something that had never been

59:55done before?

59:57And then we won Motor Trend Car of the Year.

59:59Yes, and we won Motor Trend Car of the Year in 2013.

1:00:01our first, our first, you know, great award first car.

1:00:05Pretty good start, kind of a home run, I think.

1:00:10But that whole process resulted in a linear process

1:00:13that you see on the screen.

1:00:14We designed first, then we engineered,

1:00:17and then we figured out how to manufacture it.

1:00:19Yeah, I mean, I think that's really important.

1:00:20When we were designing the car together,

1:00:22we didn't even know where we were gonna build it.

1:00:24And so we came up with maybe-

1:00:25Even no lars when we were doing it.

1:00:26That's true.

1:00:29So once we got Fremont, we were very fortunate,

1:00:31and we figured out the manufacturing solutions

1:00:33sort of like we were flying a plane

1:00:35and putting the wings and building the engine

1:00:36at the same time.

1:00:38So we knew we had to do better.

1:00:39Yeah, we knew he had to be better in order to scale.

1:00:42And as part of the master plan that you've read,

1:00:45Model 3 needed to be a smaller, more efficient

1:00:48and more affordable version of Model S,

1:00:51but it had to be equally great.

1:00:52It had to have all the things that people loved

1:00:54in their Model S or Model X,

1:00:56and would be much more affordable.

1:00:59And so we approached the process a little bit differently

1:01:03than the first time around.

1:01:04Now we had teams that we all work together.

1:01:07So we were able to combine design, engineering,

1:01:10and manufacturing process all at the same time.

1:01:14But somewhere along the way, we changed the manufacturing

1:01:18process to be fully automated.

1:01:21And so we leaned into this whole new way of manufacturing

1:01:23a car, but we had already engineered it.

1:01:26So things didn't quite go as well as planned.

1:01:29It was an amazing product, but it landed us in production hell.

1:01:34Many of us who lived through that carry those battle scars.

1:01:37It was a great idea, but it wasn't the right timing.

1:01:40Like Franz said, automating something

1:01:42that we designed to be built manually is super hard.

1:01:46And we have many, many failed examples of that

1:01:48at the Fremont factory that we ripped out.

1:01:50But some of them eventually still work.

1:01:51This is one I actually worked on with small team engineers.

1:01:54It's still running today.

1:01:55And some of the engineers that came by said we couldn't do it

1:01:58are no longer with the company, but it's running and is running faster than ever.

1:02:04So we kind of self-imposed constraints on the design when we were doing it to be built manually

1:02:10and we really didn't think about it.

1:02:12But despite all that, Model 3 is the best-selling EV ever.

1:02:15Yeah, and Model Y, which is derived from Model 3, is about to pass that.

1:02:23But we knew we had to improve the process further.

1:02:26And with Cybertruck, we designed a vehicle around a vision that actually started with

1:02:31the manufacturing process.

1:02:32And in this case, the materials dictated the design.

1:02:37Forming full-hard stainless steel isn't rocket science, but it sure isn't easy, and it limited

1:02:42the way we could do it.

1:02:43Yeah, absolutely.

1:02:45It really forced us to think about designing something in a way that you couldn't normally

1:02:50stamp panels, you couldn't form them in a traditional way.

1:02:53So you ended up with very linear bending processes

1:02:57that are just not in automotive kind of language

1:03:01of manufacturing today.

1:03:02But it actually created a very efficient process

1:03:06and one of the most dynamic designs ever, I believe,

1:03:10is definitely something that's gonna change

1:03:11the road landscape.

1:03:13Hopefully you guys saw it down there

1:03:14and you experienced it.

1:03:16It's definitely real.

1:03:17Those are real trucks.

1:03:18We're on our way to build them.

1:03:19But what that stainless steel opportunity did for us,

1:03:22It is let us rethink the factory footprint.

1:03:25We don't stamp those, that's a huge part of it.

1:03:27We don't even paint them.

1:03:28So our footprint got smaller and we started to think

1:03:31about innovative ways to take those constraints

1:03:34and make great products.

1:03:36But that constraint didn't really change

1:03:37the end result of the truck.

1:03:40It's a super dynamic truck and it has all the functionality

1:03:43you would expect out of any of the other competitive trucks.

1:03:47And the best thing about it, it's coming this year.

1:03:57So ideally, after all that, we would design, engineer,

1:04:01manufacture, and plan for automation happening together.

1:04:04It gives us the opportunity to question requirements.

1:04:07This is something that is fundamentally only available

1:04:09at Tesla, and the places I used to work

1:04:12and the top manufacturing companies in the world,

1:04:14they don't sit together.

1:04:15We are one team.

1:04:17Nowhere I know has all these teams together

1:04:19thinking about these processes from the very beginning.

1:04:22In fact, all of those engineering teams,

1:04:24manufacturing, design, automation,

1:04:25they're all in one work.

1:04:27They all report to one person.

1:04:28We can't point fingers at each other,

1:04:29so we have to solve them together,

1:04:31which is the best way to innovate.

1:04:34The traditional way of making a vehicle is this.

1:04:36You stamp it, you do build a body in white,

1:04:38you paint it, and you do final assembly.

1:04:40And what's interesting is these shops

1:04:42are dictated by the organizational structures that exist,

1:04:47and they're dictated by the boundaries

1:04:49that exist in the factories that are laid out.

1:04:51If something goes wrong in final assembly,

1:04:53you block the whole line,

1:04:54and you end up with buffering in between.

1:04:57This is at the tail end of its manufacturing optimization.

1:05:02Henry Ford first invented this assembly line in 1922.

1:05:05It's been 100 years, and it's really hard

1:05:08to make a change after 100 years.

1:05:10And when you watch it happen,

1:05:12it's actually really silly to a guy like me.

1:05:15You take all these stamp panels, you put them together,

1:05:18then you put them in a framing station,

1:05:20you build a body that looks something like a car,

1:05:23You put the doors on and then you paint them.

1:05:27Once you get the color, you take the doors off and then you start putting the interior

1:05:33inside the car.

1:05:34It comes in through the openings that already exist.

1:05:37I wish it went in like this big piece like this but there's actually people coming in

1:05:40and out of the car.

1:05:42There's awkward movements.

1:05:44Then we lift the car up.

1:05:45We put stuff from underneath it.

1:05:46We put it down.

1:05:47Then we put the seats in the car.

1:05:49And finally, we close it all out with glass and we bring those doors that went away for

1:05:53for a trip and we put them back in the car.

1:05:57Most of the time we're doing this with a big giant car

1:06:00and moving it and doing really nothing to it at all.

1:06:03What's funny though in this kind of whole process

1:06:05is that just recently Toyota just called this

1:06:08an engineering work of art.

1:06:10True.

1:06:13The Model Y.

1:06:15That was humbling, but at Tesla it's not good enough.

1:06:19If we're gonna scale the way we wanna do,

1:06:22we have to rethink manufacturing again.

1:06:24as part of the master plan, we have to make another step-changing cost.

1:06:28We started this on Model Y when we made these huge gigacastings, and we deleted hundreds

1:06:33of parts. We simplified assembly with the Model Y structural battery, where we decided

1:06:39the floor should be a part of the car. The battery is the floor. We put the seats in

1:06:44the interior on the battery, and we bring it up through a big open hole, and we assemble

1:06:48it. And this allowed us to do things in parallel, fully rethinking the process and reducing the

1:06:54final assembly line by about 10%.

1:06:57And we thought maybe we could do this other places.

1:06:59Yeah, I mean, in a way, the constraints become

1:07:03part of the solution rather than a problem.

1:07:07So when you think about what I'm trying to say,

1:07:09I really want to hammer this home.

1:07:11When you have a car that's about five meters long,

1:07:13and you have people working around it,

1:07:14like we did in Model 3, and we change that

1:07:17to this process where we take different parts of the car,

1:07:20and we can do more at the same time,

1:07:21like we did with the Model Y structural battery pack,

1:07:24What you see here is us doing that on the front part of the vehicle, or the rear part of the vehicle.

1:07:30That means we can get more people working on the car or robots, working on it at the same time.

1:07:36That means we have better operator density, less time doing nothing.

1:07:40I call that space-time efficiency.

1:07:42It has nothing to do with quantum mechanics.

1:07:44We can have that conversation later.

1:07:46But we get 44% more operator density, which means more work, less time walking back to the station,

1:07:5130% improvement in space time efficiency.

1:07:56And because we're not building it in and out of the car

1:07:58with those slow movements of those robots

1:08:00I showed you from production hell,

1:08:02when we go to automate it, it's going to be a lot easier.

1:08:06In the end, that will probably look something like this,

1:08:09where we balance parallel and series manufacturing

1:08:12in a way where we only do things that are necessary,

1:08:14with a much shorter final line blocking a lot less

1:08:17of the entire rest of the factory so we can optimize material

1:08:21flow using the best practices.

1:08:23And what that means, it's going to look something like this,

1:08:26where we build all the sides of the cars independently.

1:08:28We only paint what we need to.

1:08:31And then we assemble the parts of the car once and only once.

1:08:35We put them where they need to go.

1:08:37The interiors attach from a bottom up or a top down strategy,

1:08:42so there's more access for those robots and people.

1:08:44We aren't moving heavy objects around and doing nothing to it.

1:08:48And it means we're doing more work on the car more of the time.

1:08:51And then when we take all of these tested sombre assemblies

1:08:55and we put them together, we finally assemble the car only

1:08:58one time, putting the sides on with all of their parts

1:09:01to a front and rear that was already assembled,

1:09:03carrying the floor in with the seats,

1:09:05and finally boxing it out with the doors one time,

1:09:09just like Cybertruck.

1:09:11So in the end, you get the same car.

1:09:14But it's not going to be a model why.

1:09:17Yeah, this is not, this is a model Y for illustration,

1:09:20not the next gen vehicle.

1:09:23In the end, what does that mean?

1:09:25To increase the scale and adoption of electric vehicles

1:09:27on the orders of magnitude that we just showed you,

1:09:30we have to make constraints part of the solution.

1:09:32It leads us to greater than 40% reduction in footprint,

1:09:36which means we can build factories faster,

1:09:38with less CAPX and more output per unit.

1:09:41faster, less CapEx, more output per unit dollar.

1:09:47Zach's gonna go into more details on this later,

1:09:50but it also means through this innovation

1:09:52and some of what my other engineering colleagues

1:09:54are gonna talk to you about in the future,

1:09:56will reduce cost as much as 50%.

1:09:59This is the two for one concept you hear me and Elon

1:10:01talking about on earnings calls.

1:10:04Yeah, so I think our track record approves

1:10:07that we can deliver the best cars,

1:10:08and we deliver the best cars in spite of

1:10:11because of these constraints.

1:10:14And I'd love to really show you what I mean

1:10:17and unveil the next gen car.

1:10:20But you're gonna have to trust me on that

1:10:21until a later date.

1:10:23Just, and I promise we'll always be delivering exciting,

1:10:26compelling and desirable vehicles,

1:10:28like as we always have.

1:10:29Have we ever not?

1:10:30Yeah, we always do.

1:10:31But what else has Tesla known for?

1:10:34Speed.

1:10:36So with that, I wanna bring on Colin,

1:10:37who's in charge of making our cars go fast.

1:10:40Thank you Lars. Thank you Franz. My name is Colin Campbell and I have the real privilege

1:10:50of leading powertrain engineering here at Tesla. We make the fastest cars that you can

1:10:56buy for the money whether they're electric or gas and the Model S plaid that we're looking

1:11:01at here it has more than a thousand horsepower and pound per pound the motors in that car

1:11:06or as powerful as jet engines. Our cars are super fun people absolutely love driving in

1:11:11And the other thing that you probably all know about our power trains is that they're efficient.

1:11:16Ourcos go 25 to 30% further than other EVs in our class for the same amount of efficiency.

1:11:23But at Tesla, efficiency means more than just reducing how much energy the cars use.

1:11:29It's about how we develop, how we manufacture, how we refine, and how we scale the power train.

1:11:36Now the Model 3 and Y powertrain is a great example of this broader meaning of efficiency.

1:11:42So since we launched it back in 2017, we've continuously improved that powertrain and the factory that builds it.

1:11:49So the drive unit, the engine of the car, is 20% lighter for the same power.

1:11:54We use 25% less heavy rare earths than when we started.

1:11:59And the powertrain factory, which is behind me today, is 75% smaller and 65% cheaper than

1:12:07the one that we originally built.

1:12:10What I really want to emphasize is that we did all of this without compromise.

1:12:14Our cars are just as powerful.

1:12:16They go just as far.

1:12:18They cost the same or less.

1:12:20And the factories have the same output.

1:12:23So how did we do that?

1:12:27We did it by designing the entire vehicle and the entire factory together as one company,

1:12:33and this sets Tesla apart.

1:12:37We have small and highly capable teams, and to make a critical decision we can have the

1:12:41battery cell chemists, the mechanical engineers, the manufacturing engineers, the supply chain

1:12:46team, the automation designers, the software programmers, all in one room working together

1:12:53in real time.

1:12:55And that allows us to make decisions that are best for the whole car and to make them really

1:12:59fast.

1:13:00And that approach is unlike traditional automotive engineering, which is really fractured.

1:13:05And if you were to go buy a premium German electric car, the engineers who designed the

1:13:10drive inverter in that car, they did not work for that car company.

1:13:13They worked for a contractor.

1:13:14And it tests that we designed the entire car and the factory that builds it.

1:13:19And I want to highlight a few examples of what we've been able to do in-house thanks

1:13:24to that unique approach.

1:13:27So inside the charger of your Tesla

1:13:30are transistor packages.

1:13:31And that's at the top of your screen here.

1:13:33And every electron that moves you down the road

1:13:36flows through one of these packages.

1:13:38We designed our own custom package, which

1:13:41is what you're seeing here.

1:13:43And we can extract twice as much heat out of that package

1:13:46as well we could buy off the shelf.

1:13:47And so what does that mean?

1:13:49It means that the silicon carbide wafer that's

1:13:52inside those packages can be much smaller.

1:13:55And silicon carbide is an amazing semiconductor,

1:13:59but it's also expensive, and it's really hard to scale.

1:14:03So using less of it is a big win for us.

1:14:07And then on top of that, orchestrating all of these

1:14:09transistors and making them switch in the right ways

1:14:12is computationally extremely intensive.

1:14:15It used to require four microprocessors,

1:14:18which are shown here in the bottom left.

1:14:20We have developed our own custom microprocessor.

1:14:24It's purpose built for high power electronics.

1:14:27It's half the cost and it does in just one the job

1:14:31of all those four.

1:14:33And these are just two examples of many

1:14:35that I could use to showcase our expertise

1:14:38in high power electronics.

1:14:39And that expertise has allowed us to take the cost

1:14:42of the chargers that were in our Model S

1:14:44when we launched it in 2012, both the cost and the mass,

1:14:47and cut both of those in half.

1:14:52And even more important, the power electronics

1:14:53are central not just to our cars.

1:14:56They're also central to our superchargers,

1:14:59into our energy storage products.

1:15:01And Rebecca and Mike will be talking more about that.

1:15:04So in addition to the work that we've done in hardware,

1:15:08we've also done a lot of work in-house on software.

1:15:10So this is the drive unit for Model 3 and Y.

1:15:12And if we take a cross-section, we

1:15:13see the stator and the rotor.

1:15:15And they're responsible for the core function

1:15:17of the drive unit, which is converting electricity

1:15:19into motion.

1:15:21And our custom software lets us simulate the rotating magnetic

1:15:25field that is responsible for that conversion.

1:15:28And getting that simulation exactly right,

1:15:31it's central to the cost, the weight, the size,

1:15:35and even the sound of the drive unit.

1:15:38Now, you can buy software that will do all of this,

1:15:42but our tools are faster, and they are more accurate,

1:15:45and that was not easy to do.

1:15:48And that allows us to quickly iterate

1:15:50millions of possible drive unit designs to find the best one.

1:15:55We want to highlight one more area where Tesla really excels because we integrate work that

1:16:00is often farmed out.

1:16:02So when you are making a new product, it's not enough to think about the product itself.

1:16:07You have to think about how you're going to make it at scale.

1:16:10So Tesla, our powertrain, and our powertrain manufacturing equipment is both designed under

1:16:15one roof.

1:16:16The engineers who are designing the motor, they are in the same room as the engineers who

1:16:20who are designing the machine that's

1:16:22going to put that motor together.

1:16:24And that collaboration pushes us from day one

1:16:27to design products that are not only high performance,

1:16:30but are really easy to assemble.

1:16:33So all of this expertise that we have in the Powertrain team

1:16:38and hardware and software in manufacturing,

1:16:41it's going to have a major impact on our next platform

1:16:45in our next Powertrain.

1:16:47So the silicon carbide transistors

1:16:49that I mentioned that are key component,

1:16:51but expensive, we figured out a way to use 75% less

1:16:56without compromising the performance

1:16:57or the efficiency of the car.

1:17:00And of course, we know that battery cell supplies

1:17:02one of the constraints on the scalability of EVs right now.

1:17:06Our new powertrain is compatible with any battery chemistry.

1:17:10That will give us great flexibility in battery sourcing.

1:17:16If we want to make EVs more accessible to people,

1:17:17they have to be cheaper.

1:17:20We've reduced the driving at cost to about $1,000.

1:17:23And we don't think any other automaker is even close

1:17:26to that number.

1:17:29Finally, the bigger a factory is, the longer it takes to build.

1:17:33If we can build the same number of cars from a smaller factory,

1:17:36we are going to be able to scale EV production faster.

1:17:39Our next powertrain factory is 50% smaller

1:17:42than the one that's behind me today,

1:17:44even though it has the same capacity.

1:17:49All of these improvements, I think,

1:17:50are going to be transformative for the adoption of EVs

1:17:53and our ability to scale them, there's one more thing

1:17:55that I want to highlight.

1:17:58So I talked about how we had reduced the amount of rare earth

1:18:00in our power trains.

1:18:02And as the world transitions to clean energy,

1:18:05the demand for rare earth is really increasing dramatically.

1:18:08And not only is it going to be a little hard to meet that demand,

1:18:11but mining that rare earth, it has environmental and health

1:18:15risks.

1:18:16So we want to do even better than this.

1:18:20We have designed our next drive unit,

1:18:22which uses a permanent magnet motor

1:18:24to not use any rare earth materials at all.

1:18:29So how does all this fit into the master plan?

1:18:32We can make lower cost products

1:18:34that are still efficient and compelling

1:18:35and we can make them at scale.

1:18:37We can use less of constrained commodities,

1:18:40silicon carbide, rare earths.

1:18:42We're gonna build them all in compact

1:18:43and high output factories

1:18:44that are easy for us to build quickly.

1:18:49We're gonna make that easy to scale power trains

1:18:51all the way up to the levels

1:18:52that Drew and Elon mentioned at the beginning.

1:18:55And this achievement,

1:18:55Like all of the achievements that I mentioned today, it's only possible because of the incredible

1:19:00people on our Powertrain teams.

1:19:01They are absolutely committed to the cause of sustainable energy, and that is why we can

1:19:06do what no other company can do.

1:19:10Thank you.

1:19:18So next I'd like to welcome my colleagues Pete Bannon, David Lau.

1:19:22Thanks Colin.

1:19:27Good afternoon everybody and welcome to Texas.

1:19:31Pictured here is the low voltage system of a Model S from 2012.

1:19:35This car was primarily designed with vendor source controllers integrated by Tesla.

1:19:40There were over 300 low voltage devices that encompass everything from the mundane, like

1:19:46the light in the glove box, to the complicated, like the infotainment computer, to the safety

1:19:51critical, like the airbags, stir-by wire, and brake-by wire.

1:19:58The low voltage harness is built from individual wires cut to length, crimped, and inserted

1:20:03into connectors, a manual process that is tedious, error-prone, and doesn't scale well.

1:20:12Going forward, we want to reduce the size and complexity of the harness and enable automated

1:20:16manufacturing.

1:20:17These wire harnesses introduce extraordinary complications, especially in the early stages

1:20:22of development and bring up of a vehicle, because when we're trying to bring up this

1:20:26entire system and we see that it's not working properly, we don't know whether it's a problem

1:20:30with the software, with the controller, the processor,

1:20:33any of the endpoints in this entanglement of wires.

1:20:35And so we have to go and debug everything all at once,

1:20:38and Pete's gonna tell us about how we're gonna make it better.

1:20:42So I'm gonna talk briefly about a few things

1:20:44that we've changed over the years since the Model S.

1:20:47With the Model 3, Teggra Tesla started designing

1:20:50an increasing share of our controllers

1:20:52by emerging controllers together.

1:20:55And we were able to simplify the design

1:20:57with a significant reduction in wire count and weight.

1:21:01Going from S to three, we reduced the wire harness

1:21:04by 17 kilograms compared to Model S.

1:21:07And this is a pretty big deal,

1:21:09and just to put that into context,

1:21:10our VP of engineering Lars will deliver a bottle

1:21:13of your favorite spirit to your desk

1:21:15if you manage to save a kilogram of weight from the car.

1:21:18So this improvement cost him quite a lot.

1:21:24Do we skip one?

1:21:25Yeah.

1:21:25The original Model 3 controllers were enhanced and then built into Model Y.

1:21:30Those improved controllers were then moved back into Model 3 when we introduced the heat pump.

1:21:35At that point, we had shared controllers across both cars, which helped simplify our supply chain.

1:21:41Those Model 3 and Y controller designs were then updated and enhanced for the new versions of Model S and X that we introduced in 2021.

1:21:49For CyberTruck, we are now designing 85% of the controllers in the car.

1:21:53For the next gen platform, we're going to finish the job

1:21:55and be designing 100% of the controllers

1:21:58to give us full control of the design and the supply chain

1:22:01at the component level.

1:22:03Having full control of the supply chain at the component level

1:22:05has been critical to Tesla over the last few years

1:22:08as the supply constraints have hampered our ability

1:22:11to build cars.

1:22:13And with that control comes control over all the software,

1:22:16which enables us to develop features and functionality

1:22:20that we never even dreamed of at the time we designed

1:22:22hardware. So that's why you see software in your car is getting better and better over time in ways

1:22:27that we didn't even think of when we designed the hardware in the first place. Yes, Century Mode is

1:22:31one of my favorite last-minute changes. With the introduction of Model 3 in 2017, we deleted relays

1:22:41and fuses from the car in favor of eFuses. eFuses replace moving parts with solid state transistors.

1:22:47They provide fine-grained control of the power system to software

1:22:51and allow software to do advanced things like load shedding in the car under adverse conditions.

1:22:56Efuses also allow for software controlled retries for transient faults

1:23:01and they all allow for detailed monitoring of the power system over time.

1:23:05And what the theme you hear a lot about today is software controlled hardware.

1:23:09And when I think about that in my team, software controlled hardware is fundamentally about

1:23:13being able to sidestep what otherwise would have been static trade-offs between one attribute

1:23:18or another in a piece of hardware.

1:23:20We are able with software to instill intelligence, context awareness, and context specific behavior

1:23:28into what otherwise would have been a piece of hardware that had to get optimized for one

1:23:32type of scenario.

1:23:33We get more of everything.

1:23:35In 2022, we completed the transition of lead acid batteries to lithium ion batteries, and

1:23:41which use a new tool-less connector.

1:23:43We expect lithium ion batteries to last

1:23:45for the life of the car,

1:23:47so we don't expect anyone to wake up

1:23:49to a dead battery ever again.

1:23:51This eliminates a major source of failures for our car,

1:23:55and the tool-less connector makes it easier

1:23:58to service the car,

1:23:59and it also includes a software feature

1:24:02to allow validation that the connector

1:24:05has been properly installed at the end of a service event,

1:24:08which removes another source of failures in our cars.

1:24:11Obviously the mass and volume savings

1:24:13are also significant, which is super helpful.

1:24:19A Tesla were always trying to improve

1:24:21every single component in the car,

1:24:23and a nice example of that is the 15-inch display

1:24:25that was originally shipped in the 2017 Model 3.

1:24:29Over time, the cost of the display has gone down 24%.

1:24:34We've been able to reduce the weight by 12%,

1:24:37and we've reduced the power by 33%.

1:24:39At the same time, we've increased the brightness of the display by 50% and improved the color

1:24:43accuracy.

1:24:44So this is one of our favorite things at Tesla, just to make a component cheaper and better.

1:24:50At Tesla, the drive to improve efficiency in the car is never ending.

1:24:53I think of it as peeling a car at it, you just take a swipe at it, a little bit falls off,

1:24:58doesn't really make much of a difference.

1:25:00But if you just keep at it, over time it accumulates into a pretty significant improvement.

1:25:05One of the changes that we're looking to make right now is to change something that's been

1:25:09steady for the last 60 years, 12 volts.

1:25:1312 volts has been, as I said, for 60 years.

1:25:16The demand for power in the car has been steadily increasing

1:25:19to the point where we now have to have pretty large wires

1:25:22to drive over 200 amps of current around the car,

1:25:24which increases the mass and cost.

1:25:26With Cybertruck and all future Tesla platforms,

1:25:29we will be moving to 48 volts.

1:25:31This reduces the current needed by a factor of four.

1:25:34And since power loss in the harness

1:25:36is resistance times the square of the current.

1:25:39A 4x reduction in current leads to a 16x reduction

1:25:42in lost power while distributing energy in the car.

1:25:46That allows for smaller wires, smaller e-fuses,

1:25:49and smaller controllers.

1:25:50It also allows us to make those heat sinks smaller,

1:25:53or in many cases, remove it completely,

1:25:55all benefiting the car in terms of mass and weight.

1:26:0048 volts is the future for a little voltage design

1:26:03at Tesla, and likely the rest of the industry

1:26:05and do course.

1:26:06We welcome and encourage other OEMs

1:26:08and the entire supplier network to join us

1:26:11on this evolution.

1:26:12Mm-hmm, you bet.

1:26:16The number of wires in the car is driven

1:26:18by the number of endpoints that need to be powered and controlled.

1:26:21In the past centralized control has

1:26:23led for wires spanning the entire car.

1:26:26For Cybertruck design, we have moved to a local controller

1:26:29where the wire is connected to the nearest controller,

1:26:34and those controllers are connected over ethernet.

1:26:36Wires are routed to the nearest controller

1:26:38where the data is converted into a network packet

1:26:40for a transmission to the correct location in the car.

1:26:44To be effective, the network must be reliable,

1:26:46have low latency and low jitter,

1:26:48and these are all attributes that we've been able to achieve

1:26:50with the current design.

1:26:52The design has eliminated most of the cross-work car wires

1:26:55in Cybertruck, and with the next gen platform,

1:26:58we're gonna finish the job and eliminate all of them.

1:27:01The some consolidated vehicle network allows us also

1:27:03to make a whole lot more dynamic changes on the fly.

1:27:06to how components in the vehicle talk to each other,

1:27:09rather than the traditional approach

1:27:11of separate CAN buses that are spread

1:27:12throughout the vehicle.

1:27:13Yeah, and one of the nice things is that you can see

1:27:17the entire vehicle through a single connection,

1:27:19which wasn't possible in the past, for debug.

1:27:23For the next-gen platform, we're looking to optimize

1:27:25the controller design across the entire car,

1:27:28and across all of the organizations,

1:27:29not just for a given subsystem,

1:27:31simplifying the wire harness will enable automation.

1:27:3448 volts will allow us to reduce the size, mass, and cost of the low voltage system.

1:27:41And it's one of the key things to enabling us to scale production of the low voltage systems

1:27:46to a further master plan 3.

1:27:49David?

1:27:54So I joined Tesla in 2012.

1:27:56And even in those very early days of the Model S, we already had the ability to update software

1:28:02over the air in all of the controllers throughout the entire vehicle.

1:28:05At a time when other cars were just starting to update software in only their infotainment units.

1:28:11And furthermore, we've had sophisticated anonymized data logging and telemetry capabilities

1:28:17that allow us to understand how the fleet is doing, what customers are doing with it,

1:28:22and how it's responding in the field to that usage.

1:28:25These two capabilities combined, OTA updates and data insights,

1:28:31Give us the ability to iterate quickly on our software and to maximize the amount that we learn and proceed

1:28:38and the progress that we make in every one of those iterations.

1:28:41We've used these capabilities to inform countless design decisions in both software and hardware.

1:28:47One example is that we were able to monitor and track the use of the sunroof in our cars

1:28:54and found that our customers never use their sunroof.

1:28:56So it made easy decision to remove it.

1:28:59Very barely.

1:29:00Yeah.

1:29:01I haven't had some reps for a while.

1:29:05As another example, we are able to use our collision data to design our crash safety systems

1:29:11for what happens in the real world, far beyond what regulatory and consumer ratings tests

1:29:16prescribe.

1:29:18And then furthermore, we're able to use that data to recreate in simulation all of those

1:29:23crashes every single time we make a change to the vehicle's design.

1:29:28to the software that controls the airbags and other restraint systems.

1:29:32Here on this graph, the gray dots that show up first

1:29:36are specific crash tests that are prescribed by regulatory and consumer

1:29:40ratings agencies throughout the whole world. And the rest of the

1:29:44color that fills in around and on top of it, those represent the crashes

1:29:48that we've seen happen in the real world. Those are what we

1:29:52designed for and what we test to. Right. And last year, this data was used to

1:29:57used to change the algorithm for seat belt tensioning to reduce injuries in the field,

1:30:02which was OTA to all of our cars.

1:30:06Here are a few statistics that demonstrate how staggeringly quickly we gain new insights

1:30:11from the anonymized data that we've received from the fleet every day.

1:30:16Every day, 123 million miles driven, 1.9 million charge sessions, and those rates are only

1:30:23increasing as we deliver more and more cars into the world.

1:30:29As another example, we've analyzed patterns in the way the fleet drives and charges to

1:30:34optimally size battery packs for our next generations of vehicles.

1:30:38And you'll hear lots of other examples throughout the rest of the day about how we're using data

1:30:42to inform decisions in the product and in manufacturing.

1:30:46So you've heard a lot from Pete, Colin, and others about how important vertical integration

1:30:50is for us in hardware, but it's especially important for us in software.

1:30:56And this is extremely difficult from the way the rest of the traditional automotive supply

1:31:00chain is set up.

1:31:02In most situations, all the controllers in the car are delivered by different tier one

1:31:06suppliers whose software is written by tier two suppliers who farm some of that software

1:31:12out to different tier three suppliers and so on and so on.

1:31:17Making a change that spans multiple components takes months of coordination before any work

1:31:22can even start.

1:31:24And furthermore, integration of software in the server side with what's happening in our

1:31:29vehicles has always been a core part of our DNA and enables us to do things that nobody

1:31:33else can.

1:31:35Most engineers think about the vehicle as a complete, consolidated, fully contained system.

1:31:41But in the software team, we think about the system as including the vehicle as well as

1:31:46all of our back-end server side applications and infrastructure and the resultant feedback

1:31:51loop from the entire fleet that informs all the decisions we make as an engineering team.

1:31:57For example, we recently released a feature in Model S and X that automatically, predictably,

1:32:04adjusts and raises the suspension for ride comfort before the car hits a section of rough

1:32:09road.

1:32:10We do that by leveraging the fleet to generate a map of road roughness everywhere our cars

1:32:15drive.

1:32:16Sounds pretty simple, but requires coordination of software across a number of different

1:32:20components inside and outside of the car.

1:32:24The restraint control module, whose inertial measurement unit

1:32:26senses the road roughness, the autopilot computer,

1:32:30whose GPS module and localizer figure out

1:32:33where the car is in the world, our navigation server, which

1:32:37aggregates the anonymized telemetry from the fleet

1:32:39and annotates our map with things like updated speed limits,

1:32:43lane topology, and now road roughness.

1:32:47The onboard navigation engine, which

1:32:49looks ahead of the route that the car is traveling on

1:32:52and determines whether things are about to get rough.

1:32:55And finally, the air suspension controller,

1:32:57which takes all of those factors into account

1:33:00and decides continuously whether it

1:33:02is appropriate to adjust a suspension for ride comfort.

1:33:07And before we release this feature to the world,

1:33:10we sent prototype versions of it to the entire fleet that

1:33:14ran passively in the background, sending us anonymous data

1:33:18about every time it would have engaged,

1:33:20which taught us exactly how it was going to behave

1:33:23in the wild in every environment, in every circumstance.

1:33:27Yeah, being able to test new algorithms in the background

1:33:31without an impact in the car is really critical ability

1:33:34for us, especially for safety critical things

1:33:35like automatic emergency braking.

1:33:38And we've used it to iterate on early versions

1:33:42of our stability control algorithms,

1:33:44which were introduced to the entire model

1:33:46three fleet in 2017 with extremely high confidence,

1:33:50even on a very foundational and safety critical function.

1:33:55OK, so far you've heard me talk a lot

1:33:58about how we leverage our software systems

1:34:00to iterate quickly on the customer-facing product.

1:34:03But what I haven't talked about, and what we haven't said

1:34:05much about publicly in general, is

1:34:07how we leverage these capabilities

1:34:09in our own internal operations.

1:34:11So for example, in the early days of manufacturing model S,

1:34:15We very quickly realized that there are a lot of different ways

1:34:19you could misassemble a car.

1:34:21You could forget to plug in a wire.

1:34:23You could not fully set a connector.

1:34:25You could pull apart from the wrong bin.

1:34:28So when we designed our manufacturing processes

1:34:30for Model 3 and all of our other vehicles going forward,

1:34:34we took a page from the software playbook, which says,

1:34:37test early and test often.

1:34:41And we applied that mantra to every single car

1:34:44that we build at every step of the assembly line.

1:34:48So now, when a production associate

1:34:49plugs something into the car, the central car computer

1:34:52sees that connection, confirms that it's

1:34:55the correct part for the type of car that's

1:34:56supposed to be built, installs a software update,

1:34:59if necessary, checks configurations and calibrations,

1:35:02and runs a barrage of tests on that thing,

1:35:04and all the other things that are connected to it.

1:35:06And if it finds an anomaly that can't be fixed purely

1:35:09with software, it throws an alert immediately

1:35:11that is displayed prominently on the vehicle's display,

1:35:14as well as sent to back-end command and control systems

1:35:17so that a human can come over and fix the problem

1:35:20before the rest of the car gets built on top of it.

1:35:24And we're taking a similar approach in service

1:35:26through a combination of onboard vehicle diagnostics,

1:35:30natural language processing applied to customer narratives

1:35:33when a customer schedules a service appointment,

1:35:35and a suite of internally developed tools,

1:35:37we are successfully diagnosing, scheduling,

1:35:41and ordering parts for over 33% of customer concerns and service.

1:35:46Yep, and we also use testers built around our own controllers to test sub-assemblies

1:35:51before they get to the car, which will be even more important as we go to Lars' unboxed

1:35:55assembly strategy.

1:35:56100% will know with extremely high confidence that before those boxes get put together,

1:36:02that they are assembled, they themselves are assembled correctly.

1:36:08Okay, you've heard at lots of other events about all the awesome things that are auto-pilot

1:36:12and AI teams are doing to make the car drive itself,

1:36:16bringing us into the future of autonomy.

1:36:18But we've also been thinking for years

1:36:20about all of the other pieces that we are going to need

1:36:23to manage a network of autonomous vehicles.

1:36:27A lot of this has been happening behind the scenes

1:36:29in the form of platform level functionality

1:36:31that we'll leverage later.

1:36:33But you've seen some of it surface already

1:36:35in terms of features that our customers

1:36:37and our internal operations teams can benefit from.

1:36:40Like in 2021, we built on top of our mobile app phone key

1:36:43and gave our customers the ability to share their car

1:36:45with anyone by sending them an invitation

1:36:48in a text message or an email.

1:36:50Last year, we introduced profile synchronization,

1:36:53which synchronizes your seat, steering, and mirror positions,

1:36:56as well as your settings, media favorites,

1:36:58and stored navigation locations

1:37:00across all vehicles in your account.

1:37:03Internally, we have an app that allows our personnel

1:37:06in engineering, manufacturing, delivery, and logistics

1:37:10to view, locate, and drive all Tesla-owned vehicles

1:37:14at their site.

1:37:16And when customers bring their car into service,

1:37:19when they receive a Tesla-owned service loaner vehicle,

1:37:22we are starting at some locations to automatically

1:37:25add that vehicle to the customer's mobile app account,

1:37:27so that combined with cloud synchronized profiles

1:37:30and phone key, it's a completely seamless experience.

1:37:33As soon as the customer walks up to that loaner,

1:37:35it behaves and feels exactly like their own car.

1:37:39And all this is built on top of end-to-end encryption

1:37:43and cryptographically signed commands

1:37:46so that customer data remains private and obscure

1:37:49and the fleet only trusts commands from authorized parties.

1:37:54So that's a bit of a preview of all the pieces

1:37:56that we're building looking ahead

1:37:58to our future of autonomous fleets.

1:38:00Synchronization, permissions management,

1:38:03security and privacy.

1:38:07And that has been a whirlwind tour

1:38:09through some of the things that we're doing behind the scenes

1:38:12to enable the efficiency, cost reduction, and speed

1:38:15that we're going to need for our next phase of growth

1:38:18and for the next phase of the master plan.

1:38:20And so now, to tell us about the latest in full self-driving,

1:38:25I'd like to introduce a show.

1:38:35Thanks, David and Pete.

1:38:36Hey, everyone.

1:38:37My name is Ashok Eliswamy.

1:38:38I work on autopilot and cell driving at Tesla.

1:38:41I joined Tesla back in 2014.

1:38:43So I've been working on this problem for almost nine years

1:38:45now.

1:38:48Yeah.

1:38:48Some of you might be wondering, hey, what is this self-driving

1:38:50got to do with the plan to a sustainable future?

1:38:54But it's actually a critical part of this plan.

1:38:56And here's why.

1:38:59Currently, when the car is not being used,

1:39:01it is sitting idly in parking lots, not doing anything.

1:39:05But when autonomy is truly unlocked,

1:39:08this car, instead of being idle, can go serve other customers.

1:39:13And this fundamentally reduces the need to scale manufacturing to extreme levels, because

1:39:21each car is being used way more.

1:39:26But this is no trivial problem.

1:39:28Building a scalable self-driving system is, I think, one of the hardest real-world AI

1:39:33problems out there right now.

1:39:35Nonetheless, we at Tesla have made significant strides in making one of the most general systems

1:39:43solving this problem.

1:39:44There are three main parts to get right to build a scalable problem.

1:39:53First is the architecture.

1:39:56The architecture of the AI system.

1:39:58Second is the data, and third is the compute.

1:40:02We'll start with the architecture.

1:40:06At Tesla, we are betting on AI mission-learning neural networks to help us build a general vision and planning system.

1:40:13In the early days, we used to have single camera, single frame neural networks that produced some

1:40:20outputs.

1:40:21These were switched together in some post-processing steps for the planner.

1:40:25But this was very brittle and was not leading to great success.

1:40:28So what we did in the last few years is transition most of our stack into this multi-camera video

1:40:35neural networks.

1:40:37These neural networks take in the live feed from the car in real time of the eight cameras

1:40:43in the car and produce a single unify 3D output space.

1:40:51There are many tasks that we produce,

1:40:52such as the presence of obstacles, their motion, lanes,

1:40:56roads, traffic lights, what have you.

1:41:00This is one example of output that you're seeing here.

1:41:02This is from our occupancy network

1:41:04that predicts the positions of obstacles and their motion.

1:41:07You can see that it precisely captures

1:41:09the swirby, violent motion of this truck next to us.

1:41:13And this helps the planning system

1:41:15to avoid a collision with this object.

1:41:21Some of the tasks, such as lane connectivity,

1:41:23are more complicated to model using

1:41:25naive methods in computer vision.

1:41:27So we don't stop at the computer vision techniques,

1:41:29but we reach out to techniques in other areas,

1:41:32such as language modeling, reinforcement learning,

1:41:35to model this task.

1:41:37This is similar.

1:41:38These networks use similar techniques,

1:41:40such as transformers, attention modules,

1:41:42auto-regressive modeling of tokens,

1:41:46similar to what large language models like chat GPT do out there.

1:41:51With such an end-to-end system of solving perception,

1:41:55we have really removed the brittle post-processing steps

1:41:58and produced high quality outputs for the planning system.

1:42:03Even the planning system is not stuck in the old ways.

1:42:06It is now starting to use more and more AI systems

1:42:08to solve this problem.

1:42:11The neural network-based planners are needed,

1:42:13especially in complicated urban planning,

1:42:15when there's a lot of other objects interacting with us.

1:42:19This is, for example, an intersection

1:42:20where we had to turn left while yielding to crossing objects

1:42:23and to the pedestrians crossing the road.

1:42:25We had to do this both safely and smoothly

1:42:29while respecting everyone's right-of-way and preferences.

1:42:34If this was done naively, each configuration

1:42:37would take 10 milliseconds of compute

1:42:38and there are easily thousands of configurations

1:42:41to reason about.

1:42:43This would not be feasible using traditional compute.

1:42:46But using AI, we have packaged all of this

1:42:49into a 50-millisecat compute budget

1:42:51so it can run in real time.

1:42:56The second big piece of this puzzle is the data.

1:42:59This is where Tesla has a unique advantage

1:43:00because it can tap into the fleet to access the exact data

1:43:06that can fix the problems.

1:43:11But raw data is not sufficient.

1:43:14In order to train these networks, you need label data.

1:43:18you need labels to supervise the networks.

1:43:21And if you only depended on the human labelers,

1:43:25this data would be too tiny to train

1:43:27these large multi-camera video modules.

1:43:30We need lots and lots of data to train these networks.

1:43:34Hence, we have built a sophisticated auto-labeling

1:43:38pipeline that collects data from the fleet,

1:43:42runs computational algorithms in our data center,

1:43:45and then produces the labels to train these networks.

1:43:49Here, you are seeing a 3D reconstruction

1:43:51that is happening by collecting various clips

1:43:54from different cars in the fleet and assembling them all of them

1:43:57together into a single unified representation

1:43:59of the world around the car.

1:44:01You can see all the lanes, the road boundaries, curb,

1:44:03crosswalks, even the text on the road,

1:44:06being accurately reconstructed by these algorithms.

1:44:13We don't have to stop at this kind of reconstruction.

1:44:16Once we have this base reconstruction,

1:44:18we can build various simulations on top of it

1:44:21to produce an infinite variety of data

1:44:23to train for all the corner cases.

1:44:26We have a capable simulator that can synthesize

1:44:30adversarial weather, lighting conditions,

1:44:32and even the emotion of other objects

1:44:35to test all the corner cases that might even

1:44:37be rare in the real world.

1:44:42Here's an example of why this data is critical

1:44:44and how we can solve problems using data.

1:44:47So back in the days, for example, in this case,

1:44:51we had some false braking where we thought this car that's

1:44:54actually parked there was going to move into our path

1:44:56and hence we were precautiously braking.

1:45:00But it is unnecessary because the car is truly parked.

1:45:04So how can we solve this?

1:45:05What we did was we mined the fleet for similar cases

1:45:09where the car falls braking due to some parked car.

1:45:13We added 14,000 videos to our training set.

1:45:16And then once we train the networks again

1:45:18with this new data, it now correctly understands that,

1:45:21OK, there is no driver in this car,

1:45:23and it must be parked, so then there is no need to break.

1:45:26And this is a systematic way to solve problems.

1:45:29On the chart on the right side, you can see that every time

1:45:32we add data, the performance improves.

1:45:35And then we can do this for every kind of task

1:45:37that we have in our system.

1:45:40And this is what we call as data engine.

1:45:43We identify challenging cases, such as the one you saw earlier.

1:45:46and there could be other different types of challenging cases

1:45:48too.

1:45:49We mined the fleet for such data, put it through our auto-lapping

1:45:54system, and produced the labels, added to the training set.

1:45:58And once you have the newly trained models,

1:46:00we're deployed on to the fleet.

1:46:04If we rinse and repeat this process,

1:46:06everything gets better and better.

1:46:12The final critical piece is the compute.

1:46:15In order to train these large models

1:46:16in a reasonable amount of time, you need lots of compute.

1:46:20In addition, compute is also required

1:46:22to produce the labels automatically.

1:46:24This is just compute data centers.

1:46:26In addition, we also have high compute computers

1:46:32in the car, which can run up to 150 tera ops of compute.

1:46:37On the back end, we have a 14,000 GPU cluster,

1:46:41and roughly 30% is used for auto labeling,

1:46:43and the remaining 70% is used for training.

1:46:47We also have a 30 petabyte of video cache,

1:46:50And this is growing to 200 petabytes.

1:46:53All of this is going to significantly increase once

1:46:56we bring Dojo, which is our training computer,

1:46:59on board into this.

1:47:01Just to give you a reference, just the occupancy networks

1:47:03that you saw earlier, use 1.4 billion frames

1:47:06to train these networks.

1:47:12We have already shipped our FSD beta software to pretty much

1:47:15everyone that has bought it in United States and Canada.

1:47:19This amounts to roughly 400,000 customers

1:47:21who can turn it on anywhere.

1:47:23And then the car would attempt to drive to the destination.

1:47:25It's still supervised.

1:47:28But it can already handle turns, stop by traffic lights,

1:47:31yield with other objects, and generally get to the destination.

1:47:36We have observed that people who use FSD

1:47:38are already 5 to 6x safer than the US national average.

1:47:44Like I mentioned earlier, the solution to scalable FSD

1:47:47is getting the architecture, the data,

1:47:51and the compute just right.

1:47:53And we have assembled a world-class team to execute on this.

1:47:56We're pushing the front here on these three items.

1:47:59And as we improve the safety, the reliability,

1:48:02and the comfort of our system, we can then unlock

1:48:04rivalness operations, which then makes the car

1:48:08be used way more than what it's used right now.

1:48:13Is that a right to introduce Elon back on stage

1:48:16to give some more updates?

1:48:28Yeah, this is the next video.

1:48:37It's going to weird seeing the arms and legs just separate.

1:48:46a whole lab full of arms and legs with bearing in mind that when we did AID, this version of

1:49:05Optimus didn't work at all. So the rate of improvement here I think is quite significant.

1:49:13It's obviously not doing parkour but it is walking around and we have multiple copies

1:49:22I suppose, of Optimus.

1:49:29The thing that I think Tesla brings the table that others

1:49:34don't have is that we have the real world AI.

1:49:45We're the most advanced in real world AI.

1:49:48So the same AI that drives the car, which you can think of,

1:49:56the car really is a robot on wheels.

1:49:58and this is a robot on legs.

1:50:00So as we solve real world AI,

1:50:04I don't think there's any,

1:50:05I don't think there's anyone even close

1:50:06to Tesla on solving real world AI.

1:50:09That same computer and software

1:50:14goes into Optimus.

1:50:16So it's so that helpful to have a humanoid robot

1:50:19if you have to program every individual action.

1:50:22It needs to be able to walk around

1:50:25autonomously and solve tasks.

1:50:26You should be able to instruct it in simple things by showing visually what the robot needs

1:50:34to do or just telling it what to do.

1:50:37So I think that's the key advantage that we have.

1:50:41And then we also are good at designing things for manufacturing and then manufacturing itself.

1:50:47So the actuators in Optimus are all custom designed Tesla actuators.

1:50:53So we designed the electric motor, the gearbox, the power electronics, obviously the battery

1:50:58pack, everything else that goes into Optimus.

1:51:03We were quite surprised to find how little was available off the shelf because there's

1:51:10a lot of a vast number of electric motors, gearboxes and whatnot, that are available

1:51:16in the world.

1:51:17And we found none of them were useful in a humanoid robot, literally none.

1:51:23So you have to custom design the actuators for a human-rayed robot.

1:51:34And so the same team that designed the groundbreaking electric motors that are in the, say, the

1:51:40model is plaid, designed the actuators in the robot.

1:51:47So I mean, for practical purposes, what this means is that we should be able to bring an

1:51:52an actual product to market at scale that is useful,

1:51:57far faster than anyone else.

1:52:02And assuming the things I'm saying are true,

1:52:05or at least, I think they are true,

1:52:08it's just a question of the timing.

1:52:12You start getting into interesting questions of,

1:52:13like what's the ratio of humans to humanoid robots?

1:52:17I think it might be greater than one to one,

1:52:22because you could sort of see a home use for robots,

1:52:26certainly industrial uses for robots, human-eyed robots.

1:52:32I think we might exceed a one-to-one ratio

1:52:35of human-eyed robots to humans.

1:52:41It's not even clear what an economy means at that point.

1:52:44Since an economy is output per person times persons,

1:52:49but if output is much higher

1:52:50and there's no limit on persons,

1:52:53then what's the actual limit on the economy?

1:52:59We're still pretty far from carto-shift scales here,

1:53:02but we're getting there.

1:53:06So anyway, so probably the least understood

1:53:11or appreciated part of what we're doing at Tesla,

1:53:14but will probably be worth significantly more

1:53:16than the car side of things long term.

1:53:22So charging.

1:53:22I think Rebecca's next board, talking more charging.

1:53:27Hi there. My name is Rebecca Tenucci and I lead our charging infrastructure teams here at Tesla.

1:53:49And at Tesla charging, we have understood since day one that a great charging experience is the

1:53:54linchpin to electric vehicle adoption. And that understanding has meant that we've always taken

1:53:58a holistic approach to charging. That's a word that you've heard a lot here today, but what it

1:54:04means for charging is that we've considered every use case. We think a lot about what it means to

1:54:08to charge at home, even if that home is an apartment or condo.

1:54:11And we spend a lot of time thinking about,

1:54:14what does it mean to charge away from home,

1:54:15including if that's for daily commuting

1:54:17or if you're going on a road trip?

1:54:19And this holistic approach has led

1:54:21to some pretty incredible results.

1:54:23In 2022, we provided nine terawatt hours of charging

1:54:27across our various charging methods.

1:54:29Over 50% of that was supplied via convenient AC home charging.

1:54:35And when our customers are away from home,

1:54:36can visit one of our 80,000 charging points. That includes 40,000 of our beloved superchargers.

1:54:42Now, getting here has meant that we've spent 10 years building charging infrastructure

1:54:48when basically no one else in the industry would do it. And while we certainly have a

1:54:52lot of areas that we want to improve, those 10 years have afforded us the opportunity to

1:54:56get pretty good at charging. First, we have the industry's lowest deployment costs. Our

1:55:02Our costs are often 20%, if not 70%, lower than alternatives.

1:55:08And that goes for both our supercharging hardware and deployments and our AC charging product

1:55:12lines.

1:55:14A lot of reasons that we're able to achieve this.

1:55:16You've heard a lot of them here today.

1:55:17We're vertically integrated.

1:55:18We manufacture and engineer all of our own charging equipment.

1:55:21We share components across our different product lines.

1:55:25And on the supercharger side, we also install and operate all of our own sites.

1:55:31And that's led us to be pretty obsessed with finding new ways to innovate around installation.

1:55:37As an example, we have recently extended our excellence in manufacturing to how we build

1:55:42supercharger sites.

1:55:44We are pre-building four post-supercharger units at our factory in New York.

1:55:50We load them on a truck, we truck them to site, and then we crane them into position.

1:55:55This method saves us 15% on our deployment costs, and we can install a site in a matter of days.

1:56:03Once we install a site, we also operate it really efficiently.

1:56:07Over the last few years, we've been able to cut our per kilowatt

1:56:11hour cost by 40%.

1:56:13Again, a lot of reasons for this.

1:56:15But one at the top of the list is we've

1:56:16focused on increasing our site utilization.

1:56:19Site utilization is just how many sessions or kilowatt hours

1:56:22can we push through a site or a post?

1:56:25And basically, doing that allows us

1:56:27to spread our cost over more sessions, thus lower cost

1:56:30per kilowatt hour.

1:56:31But of course, that's easier said than done

1:56:33because when we push up site efficiency, of course, the risk

1:56:36is that we have a poor customer experience

1:56:39and we have wait time at our sites.

1:56:42This is where Trip Planner comes in.

1:56:44Trip Planners are in vehicle routing or navigation system.

1:56:48Other electric vehicle manufacturers,

1:56:50well, some of them have vehicle-side data.

1:56:53Other infrastructure providers have site data,

1:56:56but at Tesla, we have both.

1:56:57And what that allows us to do is to use Trip Planner

1:57:00to route vehicles towards available sites

1:57:03away from congested sites so we can balance our utilization without risking wait time.

1:57:08And the results speak for themselves.

1:57:10Over the last few years we've been able to drive up site utilization by 30 percent.

1:57:14That means lower per kilowatt hour costs while also cutting our wait time in half.

1:57:20And we think this can get even better.

1:57:21Today we're feeding trip planner with real time site information and information about

1:57:25vehicles currently at sites.

1:57:28Going forward we'll be moving that to projecting site occupancy based on the understanding of

1:57:33what vehicles are currently routing to those sites.

1:57:35And ultimately, our vision for trip planner

1:57:38is that it's the air traffic controller

1:57:40for electric vehicle charging across all infrastructure

1:57:43on a global basis.

1:57:46And the last thing we focused on to get here

1:57:48is quicker charge times.

1:57:50We're really proud of what we've been able to accomplish here.

1:57:52We've shaved off 30% of our average supercharging site time,

1:57:57visit time, over the last few years.

1:58:00This has taken improvements on the hardware side,

1:58:03with our software, on our vehicles, and on our infrastructure,

1:58:06and we're really excited to continue pushing this trend down.

1:58:11Looking forward, the job does not get any easier.

1:58:16As you heard earlier, we aspire to a fully electrified global fleet.

1:58:20That fleet from an industry standpoint

1:58:23needs nine petawatt hours of charging on an annual basis.

1:58:28And while Tesla charging certainly

1:58:29doesn't have to supply all of that,

1:58:31It does require that for our part,

1:58:33we have a few new focuses in order to scale.

1:58:37First, if you want a fully electrified global fleet,

1:58:41all of those vehicles have to have a great

1:58:43and reliable charging experience.

1:58:45So as many of you know, recently we have started opening up

1:58:50our networks on a global basis.

1:58:52Over 50% of our superchargers in Europe

1:58:55are currently open to other electric vehicles.

1:58:58We've also opened up in Asia Pacific

1:59:00with our first sites in Australia.

1:59:02And just yesterday, we opened our first 10 supercharger sites

1:59:06here in North America and the US to other electric vehicles.

1:59:10We've also invested a lot to be able to make this a really easy

1:59:14experience for our new customers.

1:59:16All you have to do is sign into our Tesla app to unlock a post

1:59:19and start charging.

1:59:20We've also, on physical site changes,

1:59:23we've also added hardware to our sites.

1:59:24They're called Magic Docs.

1:59:26And basically, that is installed in areas

1:59:28where we have different charging standards.

1:59:30and they allow for other electric vehicles

1:59:32to come to our site without having

1:59:33to bring their own adapters or hardware.

1:59:36We've also just started installing our fourth generation

1:59:39supercharger posts.

1:59:40Those are being installed in Europe first.

1:59:42And while it's not a big mention, not a big thing to mention,

1:59:46they do have longer charging cables

1:59:47so that we can more easily reach the charge ports

1:59:49of different vehicles.

1:59:52The second thing we need to make sure we do

1:59:54when we talk about scaling all of this charging infrastructure

1:59:57is we need to make sure that it's powered from renewable sources.

2:00:02We're very proud that over the last two years,

2:00:04we've procured enough renewable energy

2:00:06to offset the amount of charging we've provided to customers.

2:00:10But as we look forward and we talk about this fully electrified

2:00:13fleet on a global basis, we really

2:00:15want to make sure that the demand for charging more closely

2:00:18follows when renewable sources are available.

2:00:21Now, this chart is on an aggregate basis for the US,

2:00:24and it varies based on what your renewable generation

2:00:26sources are locally.

2:00:28But basically, what this shows is we

2:00:30want to move that charging curve and for solar and wind

2:00:34in this example, it means more daytime charging.

2:00:38And we think the best way to go towards daytime charging

2:00:41is to install AC charging that is convenient and low cost.

2:00:47Everywhere vehicles are typically parked throughout the day.

2:00:50So that's really the plan.

2:00:51While supercharging and home charging

2:00:53definitely stay a big part of the puzzle,

2:00:55Our teams are currently scaling to install AC charging

2:00:59everywhere vehicles charged during the day

2:01:00so we can power them from renewable sources.

2:01:06So summing it up, while we've spent 10 years

2:01:09installing infrastructure that is low cost,

2:01:11it's efficient and it provides a great customer experience,

2:01:14we're really just getting started

2:01:16when we talk about supporting Master Plan Part 3.

2:01:19We've got a scale capacity on an industry-wide basis

2:01:22to nine petawatt hours on an annual basis.

2:01:26We've got to open up to non-tesals

2:01:27and make the charging experience for everybody really great.

2:01:30And we've got to power all of this via renewable sources.

2:01:36And one more thing.

2:01:37Yes, we've got to scale our infrastructure.

2:01:41And yes, we want to power it via renewable sources,

2:01:44but we're Tesla.

2:01:45So we also want to make sure that we're

2:01:46continuing to focus on providing really incredible charging

2:01:50experiences.

2:01:52With that, I want to welcome up our leaders of supply chain,

2:01:55Rishan and Khan to talk a little bit about that topic.

2:02:07Thanks, Rebecca.

2:02:13Hi, my name is Karn, and my team leads supply chain

2:02:17for electronics, powertrain, and battery at Tesla.

2:02:21And we've also got responsibility for indirect purchasing,

2:02:24construction procurement, and warehousing and distribution.

2:02:27Hi, my name is Roshan.

2:02:29I've been with Tesla for over 14 years.

2:02:32My team manages vehicle commodities, solar, logistics,

2:02:37planning, and capital equipment.

2:02:3910 years, forgot that.

2:02:42So today we wanted to provide you a quick overview

2:02:45of Tesla supply chain and what really separates us

2:02:47from the typical automotive supply chain.

2:02:50As you've seen in presentations prior from Pete

2:02:53and Drew and some others, Tesla designs a lot

2:02:56of the subcomponents that go into our vehicles.

2:02:58So we're not buying things that are off the shelf

2:03:01from suppliers.

2:03:02So supply chain is not purely a commercial relationship

2:03:04as it is in most other companies.

2:03:07We actually have an arm within supply chain

2:03:09called supplier industrialization engineering.

2:03:11And the burden of taking a drawing, a design from concept,

2:03:16and turning it into thousands of products produced

2:03:18at the right cost and at the right yield,

2:03:21falls with this group.

2:03:22Of course, they work very closely with design engineering.

2:03:24But really, the level of detail that our engineers

2:03:27get involved in terms of building the capability

2:03:29of the suppliers is pretty detailed.

2:03:32So once we get a new part, naturally,

2:03:37with electric vehicles, a lot of the supply chain

2:03:38doesn't exist.

2:03:40Our engineers would basically take the drawing,

2:03:42turn it into the manufacturing concept,

2:03:44do the equipment selection, and physically go to the supplier

2:03:48and stay there for weeks or months.

2:03:49However long it takes to basically do the line bring up.

2:03:52Once the line is brought up and we're hitting the right rate,

2:03:56the right quality at the right yield,

2:03:58then we work on things like automation and yield improvement

2:04:01and those types of activities.

2:04:04So there's a big group of mechanical, electrical,

2:04:07industrial, you name it, engineers, within supply chain, and it's their responsibility

2:04:11to take care of this.

2:04:13Now this is a huge strategic advantage because we manage every detail of our supply chain

2:04:19because we're so integrated with our partners that we know about issues before they happen.

2:04:24It's almost live.

2:04:25We've got dashboards, we've got connections, and across the 10,000 or so factories across

2:04:30the world that are making our components, we have a pretty good view of the health

2:04:35of those suppliers. And this has been a key asset in how we've been able to kind of manage

2:04:39things through a lot of different issues over the past couple of years. So, you know, supply

2:04:47chain is a game of perfection. Perfection is a passing grade. It's like a, you know, if

2:04:52one part doesn't show up, that line goes down and our CEO finds out about it in less than

2:04:5720 minutes. Tesla is very different than Apple. They outsource manufacturing, we do it in-house.

2:05:02the supply chain is kind of right in the middle.

2:05:04And perfection is increasingly elusive and very expensive

2:05:09if you kind of look at all the issues

2:05:13we've had to encounter over the past year.

2:05:15And today we're just gonna talk about

2:05:17how we got through the past

2:05:19and then what are we gonna do going forward

2:05:21and how are we gonna get there.

2:05:23So let me draw your attention to this slide.

2:05:26On the left you will see a portfolio,

2:05:29or existing portfolio of products, right?

2:05:32at the site you've got the SNX platform,

2:05:34in the middle you've got three and Y,

2:05:36I think everyone's more than familiar with those,

2:05:39and then the Tesla Energy platform,

2:05:41which is a very high rate of growth for us.

2:05:45And you see-

2:05:46Don't forget the solar and superchargers.

2:05:47And solar and supercharger, thank you, Roshan.

2:05:50And you'll see a reduction in,

2:05:54and then we've basically got the tier one and tier two parts,

2:05:56this might be boring, but tier one supplier

2:06:00is any supplier that builds and supplies parts directly

2:06:03for consumption in our five factories.

2:06:05And our tier two supplier basically does the same thing

2:06:08for a tier one supplier.

2:06:09So they are the supplier to our supplier.

2:06:11And depending on the supply chain of any one of these products,

2:06:15this can actually go down to the tier six levels.

2:06:17For example, if you're talking about the battery cell,

2:06:19we really get involved with the minds of like,

2:06:22where we're procuring the lithium, the cobalt,

2:06:25and all that sort of stuff

2:06:26because we're procuring such high volumes of it.

2:06:28So you'll see a reduction in the number of parts on the SNX

2:06:32platform to the 3&Y platform.

2:06:35And this is because of the great work our design engineers

2:06:38have done of making the car simpler and easier to manufacture.

2:06:41In parallel, we actually had a similar strategy in supply

2:06:46chain.

2:06:46We had cast a wide net to a lot of supplies

2:06:48for the SNX platform.

2:06:49There's a lot of supplies that we invited to partner with us.

2:06:53And we used that platform as sort of a filter

2:06:58to see which suppliers have the technical, financial,

2:07:01and cultural capabilities to match us, right?

2:07:05Like who do we invite to the party

2:07:06when we really scale things up?

2:07:08And we've really dramatically reduced the number

2:07:11of suppliers that we were dealing with

2:07:13for the 3NY platform, and this is an approach we'll continue

2:07:16to do for future platforms as they come along,

2:07:18that are even more high volume.

2:07:20So this has really been our approach.

2:07:22It's made things a lot easier.

2:07:25And we could figure out the suppliers

2:07:29that were capable of moving at our pace,

2:07:31and then the ones that didn't have the desire

2:07:33to order the capability.

2:07:34You need both desire and capability.

2:07:37Secondly, the complexity on the tier 2 level

2:07:41is very high.

2:07:43These are the suppliers to our suppliers.

2:07:45And there are a lot of components.

2:07:47Think about silica and think about resistors, capacitors,

2:07:50diodes, all the little assemblies.

2:07:52of the raw materials.

2:07:53Raw materials, that's even further down.

2:07:56But all these little sub-assemblies

2:07:57that have to come into Tesla,

2:07:59the complexity is immense.

2:08:01So even though at the tier one level,

2:08:03we're talking about 8,000 parts total,

2:08:05it seems like a lot, but it's not a lot.

2:08:07The management of the tier two is really where we excel.

2:08:11And I'd like to illustrate that with an example.

2:08:15Meet the car computer, very innocent sounding name,

2:08:18and it's anything but.

2:08:19This is an absolute monster to manage.

2:08:23So I think you've all seen the picture on the left

2:08:25of the autopilot board.

2:08:26That's the top of the autopilot board.

2:08:28The autopilot board also has a bottom

2:08:30and the bottom is populated heavily with components.

2:08:34On the other side of the heatsink,

2:08:36you've got the MCU, the multimedia cluster board.

2:08:39This board is equally complex.

2:08:40It's also double-sided, eight-layered.

2:08:42And these boards, these computers run so hot

2:08:45at peak operation that they have to be liquid cooled

2:08:48through a heatsink.

2:08:49So this assembly requires taking those two boards

2:08:52and then bonding them to a heat sink

2:08:55that's hermetically sealed as an assembly,

2:08:58of course flashing the software and all that sort of stuff.

2:09:00And then that's one part number that comes to Tesla.

2:09:03So this is an example of one tier one part number

2:09:07that's a very complex assembly to manage

2:09:09at the tier to the level.

2:09:10And there's more than 7,000 components here.

2:09:14There's a, you know, as we stand here,

2:09:16a component's being assembled onto a car computer,

2:09:18every 1.4 milliseconds, the line that builds this computer is the length of a football field.

2:09:25So it's quite complicated.

2:09:26And initially, when we first started building this board, due to the complexity of it, we

2:09:30had to rely heavily on labor.

2:09:31But once we dialed in the quality, the rates, and the yield, we started focusing on making

2:09:36this more efficient.

2:09:37The only way to control cost is by removing labor.

2:09:40The first step is removing labor.

2:09:41The second step is fully automating.

2:09:43The third step is turning off the lights and letting the factory run, ideally.

2:09:47And that's going to be the goal for us here.

2:09:48So 95% but our work is not done,

2:09:50we're gonna be going a lot further than this.

2:09:53So this illustrates the point that a part is not a part,

2:09:57a part has a lot of complexity underneath it,

2:09:59supply chain is a game of multiple tiers.

2:10:02And what's made us successful is our involvement

2:10:05in all the details with our supply engineers

2:10:08that are Tesla employees, they're Tesla supply engineers.

2:10:10So they're like on our payroll that go and ramp up

2:10:12these capabilities that are suppliers

2:10:14and then just managing each and every attribute of it.

2:10:17You know, 7,000 a day through the chip shortage,

2:10:20through the pandemic, through all the other stuff

2:10:22that we had to deal with, was very difficult to manage.

2:10:25But because we had all the details,

2:10:27we were able to pull this off.

2:10:29Of course, supply chain is not just about parts.

2:10:32It's also about logistics and a lot of other things.

2:10:34So I'll hand it off to Roshan for sharing some stats.

2:10:39Yeah, so talking about in-bought complexity.

2:10:42Just over the course of just last year,

2:10:45We moved about 16 million pallets from our suppliers to our factories.

2:10:51Just to visualize that, if you put those pallets side by side,

2:10:55it could cover the half the circumference of the earth.

2:11:00Talking about again, the tier 2 complexity,

2:11:04over 1 billion electronic components get moved every week

2:11:09to support the production that current team manages.

2:11:12This is right now, 1 billion components a week,

2:11:1422 billion a year.

2:11:15This is going to go up, by the way.

2:11:17And on top of that, we also have the responsibility to make sure the right component is supplied

2:11:23to the right service location to support the growing fleet of our vehicles.

2:11:29And we have about 685 service centers that these components need to show up just in time.

2:11:37And then we've also been, as a result of pandemic and even before that, we've been working

2:11:44working on dual sourcing and triple sourcing some of our critical components.

2:11:49As a result of it, we have suppliers around 45 countries in the world.

2:11:53Yeah, approach really has been to bring the manufacturing of

2:11:57subcomponents closer to the point of consumption. It makes the supply chain

2:12:00more green because you're not burning diesel movement stuff around.

2:12:02And it also gives us the security of supply of having multiple sources in case one

2:12:06factory gets taken offline. And that's how we go through pandemicals.

2:12:09Yep. So last three years were incredibly hard.

2:12:16Karnan and I used to look a lot longer pre-COVID.

2:12:21But on a serious note, pandemic really changed the game of supply chain management.

2:12:28We were just not being a regular supply chain person. We were rolling up our sleeves and negotiating

2:12:34with state governments, city managers, mayors to help kickstart our suppliers to restart their

2:12:42operations.

2:12:44There are some really inspiring stories

2:12:46where our supply chain team really fought hard

2:12:50to make sure that the factories don't go down.

2:12:53And it's also a testament to the resilience

2:12:57of our supplier partners.

2:12:59Having said that, it was not all about resilience

2:13:04and never give up attitude.

2:13:07We had also laid the foundation of having dual sources.

2:13:12And also, in many instances,

2:13:15we had many things that we had simplified,

2:13:25the parts that we had simplified,

2:13:27the supply chain design that we had simplified,

2:13:29due to which we were able to get through the pandemic

2:13:32in much better shape than rest of the automotive.

2:13:35forced us to adapt. You know, we encountered issues one at a time. It was really a team effort. We

2:13:41heavily relied on the government relations team. You literally had to, the COVID pandemic, as you're

2:13:47aware, you know, happened at different countries at different times. But unfortunately, you need

2:13:50all the parts to make the car. So at any given point, there was one factory that was lying down,

2:13:54there was one jurisdiction that was being, you know, that didn't want to, they wanted to shut down

2:13:58a factory because they hadn't, you had an outbreak. So learning the rules, understanding, coming up

2:14:02up with argument points, leveraging the team and getting those. It was a difficult but

2:14:07good experience. Exactly. We learned a lot. Yeah, and as we were just figuring out to

2:14:11work with COVID, then the port congestion hit on the west coast. Logist six lead time

2:14:16went up by twice the amount and we were operating with less than 35% accuracy in the system.

2:14:24That means at any given point, the ETAs of the vessels were 75% off.

2:14:31So we figured out how to work with that and then the chip shortage hits.

2:14:36Yeah, the chip shortage.

2:14:37That was a nightmare.

2:14:40But we got through it.

2:14:41I mean, I think it's a testament to the tenacity of the team and the resilience of our supplier

2:14:46partners.

2:14:48It's mostly behind us and I think the results speak for themselves.

2:14:51Tesla was able to grow production volumes at a time where the rest of the industry had a hard time staying flat.

2:14:57I think they shrunk actually

2:14:59cumulatively

2:15:00And you know we're often asked the question how do we manage through it and and the answer is very much kind of baked through the presentations today

2:15:07First we design our own electronics. So in order to control cost

2:15:10We've really had to go and build relationships with all the tier two suppliers to negotiate them directly

2:15:16So these relationships are formed, you know, over 10 years ago, and we've really gotten to know them well.

2:15:22So when a shortage hit, we knew, you know, who to escalate to who to call, who to ask for help, and

2:15:28really kind of swarming the issue at the point of constraint.

2:15:31So the best team, you know, strong relationships with vendors, and then we also have a pretty sophisticated

2:15:38capacity planning function.

2:15:39Tier 2 capacity needs to come on earlier than tier 1 capacity. Tier 3 capacity has to be even earlier.

2:15:45And our team does a very good job of having that strategic alignment with all our partners to make sure that the capacity is ready before we need it.

2:15:53We're a mission-based company. Our goal is to put as many EVs out there.

2:15:56We did not let the pandemic phase us. We were full throttle and it was pedal to the metal in terms of our demand signal.

2:16:02That also helped the the silicon industry has very long lead times. They invest billions of dollars in

2:16:08sophisticated fabs that take years to build.

2:16:10So that stability and demand is something they appreciated and it's something that is very very much a Tesla thing

2:16:18So this was looking back, right? This is how we got here. This is how we got from zero to 40,000 cars a week

2:16:25Let's talk about how we get to 20 million

2:16:28So the question gets asked, you know

2:16:30Is Tesla gonna basically break the chip industry if he growed his 20 million vehicles the answer is no

2:16:36But let me walk you through a logic on that

2:16:38So as it stands right now, a Tesla enabled with a full self-driving computer has four times

2:16:44the silicon content of a regular vehicle.

2:16:47So at about the 2 million run rate, we're consuming about 700,000 12 inch wafer equivalents.

2:16:53That's the industry term that's thrown around for capacity in silicon.

2:16:56That's how much we're using.

2:16:57So a little below a million.

2:16:59And the global wafer capacity based on the estimates we've seen is about 135 million.

2:17:04So we're about half a percent off the industry.

2:17:06a lot. Now if you get to 20 million, when we get to 20 million, and if you don't make

2:17:11any simplifications in terms of PEP part count, which is also a separate work stream, I think

2:17:16Colin talked about it, Pete talked about it, we're not going to have all these different

2:17:19components, we're really going to simplify the architecture because simpler is better.

2:17:23But even if you don't do that, and we take the most pessimistic case, we're going to need

2:17:278 million wafers, okay? And the industry is going to scale to 200 to 200 million by

2:17:31that point. So still we're less than 5% of the market share. And if we've run the same

2:17:36offense we ran all along and continue to work with our partners and you know

2:17:40invest in fabs and invest in capacity we'll have the capacity ready by the time

2:17:45we need it so I really don't see this as an impediment from a supply chain

2:17:49perspective to get to 20 million. Now I would like to talk about a page that we

2:17:59took out of the electronic playbook and applied it to the mechanical side. Our

2:18:03vehicles use currently the heat pump based technology for the thermal system

2:18:08But before heat pump, our vehicles used to have a dispersed thermal system connected by

2:18:13hoses.

2:18:15Engineering was inspired from circuit boards and then created a system which was extremely

2:18:22integrated where 60% of the system was the size of a basketball site, basketball where

2:18:30it contained about 100 components, 50 ceiling interfaces and several different manufacturing

2:18:37processes all were packaged into this tight size.

2:18:42Now why are we talking about this?

2:18:45Well, we launched the system right

2:18:48during the outbreak of the pandemic in Model Y.

2:18:51The team was forced to manage 100 plus components

2:18:55from different suppliers around the world remotely.

2:19:01So there are four videos playing at the same time.

2:19:05One is the video of the manual line that was created

2:19:09when we started the production.

2:19:11And this manual line produced a lot of quality issues

2:19:15let alone throughput.

2:19:17So we quickly pivoted to a semi-automated line.

2:19:22But in spite of that, although the throughput was great,

2:19:27the quality issues sustained.

2:19:30So our team decided that fully automated

2:19:33is the only way to go.

2:19:36So before we automate the line, we created a 3D simulation.

2:19:42Now, not many companies or teams have this capability,

2:19:47where at this point, now when a part gets spec'd,

2:19:51we have the ability to create a 3D simulation of all

2:19:55the complex subassemblies that would

2:19:57be involved in making the part and tell the suppliers exactly

2:20:02what equipment to purchase, what to test,

2:20:04and create an end-to-end line in simulation.

2:20:08And that's what we did here.

2:20:10So with the help of a simulation, we

2:20:13had a very detailed subassembly cell, which

2:20:16was juxtaposed with the supplier's factory layout.

2:20:22And then we, after a back and forth of iteration

2:20:27and optimizing the processes, we implemented this fully

2:20:31automated line in less than eight months.

2:20:35And the results were remarkable.

2:20:3999% reduction in labor force.

2:20:44So an align that would need 1,000 associates

2:20:47to build the super manifold in full volume now just

2:20:51takes 10 associates.

2:20:55And the quality also went up by an order of magnitude.

2:20:59We have a defect rate of less than 0.05%

2:21:03and a supermanifold rolls out every seven seconds

2:21:07because we have created many automated lines.

2:21:10And these factories are the size

2:21:13of two football fields, by the way.

2:21:16Russian had to one up me on that one.

2:21:1799% two football fields.

2:21:21So why are we talking about this?

2:21:24In a world where labor costs are increasing,

2:21:28Labor participation is decreasing.

2:21:30There is a huge turnover.

2:21:33And the cost of implementing automation

2:21:36is going down by 3x in the next 10 years.

2:21:39This is how we are thinking about complex assemblies.

2:21:43And we are going to push this type of thinking

2:21:45into our supply base more and more who

2:21:48will be building our complex assemblies.

2:21:50This is the only way we think we can confidently scale

2:21:54to the 20 million vehicles target.

2:21:59To summarize, we have laid the foundation in getting ready to execute Master Plan 3.

2:22:06Our strategy is supply chain design simplification.

2:22:11We are going to make sure that we have more control into the tiers of the supply chain.

2:22:17We are going to grow responsibly and sustainably with our long-term partners and we'll automate.

2:22:24Yeah, it's right. You know, going from zero to 40,000 cars a week was tough. I mean, it

2:22:30took a lot of trial and error. It took a lot of learning. There was a lot of pain. There's

2:22:34a lot of mental scar tissue, as Elon calls it, through that process. But now that we're

2:22:39at that level, then the foundation is set. Going from 40 to 400 is not gonna, it doesn't

2:22:44really phase us. We have a capable team. We've got capable external partners that have,

2:22:48you know, gone through hell with this. Leadership team is galvanized behind the mission, and we're

2:22:53we're going to get it done.

2:22:54So I'd like to thank you for your time

2:22:56and hand it off to Drew and Tom for the next presentation.

2:23:00Thank you.

2:23:08Hi, everyone.

2:23:09My name is Tom Drew.

2:23:11I'm now responsible for global production

2:23:13at Sales and Delibert Service.

2:23:15I joined the company back in 2014.

2:23:17And I've been running the company's business

2:23:19in greater China and APAC.

2:23:22Today, I'm representing all the Gigafactories

2:23:24and talk about how we can make more cars and faster.

2:23:30So right now, we have four vehicle factories on three continents to serve our markets.

2:23:37The total number of manufacturing employees is over 65,000.

2:23:41This is also including our amazing team in Reno.

2:23:43They are making power chains, battery packs and drive unit for our vehicle factories.

2:23:48The total installed capacity in all these vehicle factories, about 2 million cars a year.

2:23:53That said, we're always looking for opportunity to grow more capacity from the existing footprint.

2:23:58So you expect this number will grow over time.

2:24:04To build more factories is the start to build more cars.

2:24:08And we certainly the expert of it.

2:24:11This is before and after image showing you

2:24:14the progress we made back in 2019 when

2:24:16we built a gig factory in Shanghai.

2:24:19So in nine and a half months time,

2:24:22we built a greenfield project to establish the vehicle factory.

2:24:27And the three months later, we start

2:24:28deliver our first customer vehicle.

2:24:32People are asking me how Tesla can build a factory that fast.

2:24:36Really, we learned from our free manufacturer a lot.

2:24:40We talked to the survivor from the production hell

2:24:43and tried to avoid all the mistakes we made.

2:24:47And we decided to design a straight line

2:24:49with the minimum number of up and downs and turns

2:24:52for easier manufacturing ability and easier construction.

2:24:55We also challenged all the assumptions

2:24:56people ever know the build a vehicle factory.

2:24:59When we delete and simplified all the redundancies

2:25:03and buffers, that's helped us to save a lot of time.

2:25:07Also, we have a very strong in-house construction team

2:25:12in Tesla.

2:25:13If there ever a job cannot be done by others better,

2:25:17and we bring this in-house.

2:25:18So we have this in-house construction team

2:25:21have a full control over all the activities on sites

2:25:24from design engineering to construction, site management.

2:25:29So this team not didn't just build Giga Shanghai,

2:25:32they also built Giga Berlin, Giga Texas, and Giga Nevada.

2:25:36And they're really the hero behind the scenes.

2:25:41Like Lars shared earlier, going forward

2:25:44with the new platform, the more modular design

2:25:48and densified Giga factory will be

2:25:51to make more cars from even reduce the footprint.

2:25:54That also means we can build more factories at the same time.

2:25:58With the joint effort of all manufacturing teams around the globe,

2:26:04I'm happy to share early this morning with the 4 million mark

2:26:08for total Tesla ever built.

2:26:14And the 4 million-some vehicle actually built in this factory,

2:26:20when you took a tour on our shop floor, you probably worked past it.

2:26:24All right, so took us 12 years to build the first million vehicle.

2:26:29And about 18 months to build the second million,

2:26:34the third million took us 11 months.

2:26:36And just a shortly less than seven months

2:26:39we'll build the four million cars.

2:26:41So we're getting better at a faster exponential growth.

2:26:46Really cool to the team.

2:26:49So what it takes to ramp a gig of factory?

2:26:52Well, if you have 600 robots, 10,000 training employees,

2:26:56or 5,000 human and 5,000 optimists,

2:27:00and hundreds of processes, you can do it.

2:27:03Sounds simple, but it's extremely hard.

2:27:05So there's two key metrics that we predominantly focus on.

2:27:09It is the overall equipment effectiveness in the cycle time.

2:27:14In Tesla, we're setting the passing

2:27:16rate for our vehicle factories with 90% OEE and 45 second

2:27:21cycle time.

2:27:22What that means, the OEE really evaluates

2:27:26the equipment uptime, the machine performance and quality.

2:27:30Simply put, this is the actual production time

2:27:35on a good quality product versus the planned

2:27:39productive time, the higher the better.

2:27:4145 second cycle time, that means you expect every 45 seconds

2:27:47as a car running off the final assembly line in the factory.

2:27:52And the faster we rent, the faster we can get to the economy

2:27:56and scale.

2:27:57If you look at the chart on the left, Shanghai,

2:28:01be able to significantly drop our labor hours per car

2:28:05during the rent.

2:28:07The little dip there is a happen in the last Q2 2022

2:28:11because of the COVID shutdown.

2:28:14And on the right, it's a Fremont motorway shop.

2:28:17Even this is a 60-year-old facility,

2:28:20the team there still will be able to optimize the material flow,

2:28:22eliminate all the signal point of failure,

2:28:25and drive higher output, hence reduce the labor hours.

2:28:30And actually, this factory keeps sending a new record.

2:28:34Yesterday, they just had a new factory daily record.

2:28:38Congrats, Fremont team.

2:28:39Absolutely.

2:28:44So how can we reduce the cycle time,

2:28:49Keep improving efficiency.

2:28:50We follow the philosophy, Elon shared,

2:28:54about a building rocket, which is questioning,

2:28:57basically find the right problem to solve.

2:28:59And we start with delete, simplify, then we

2:29:03try to accelerate, pressurize the line,

2:29:05and find the solution that actually work.

2:29:08And at last, we think about automation.

2:29:12There's one example in Gigafactory Shanghai in the paint job.

2:29:16we find that there's overlap baking range

2:29:20between the PVC sealer and the top coat.

2:29:25It was done by two different ovens,

2:29:27and we decided to combine the two processes.

2:29:30Eventually, that helped us,

2:29:32didn't just help us to reduce the cycle time,

2:29:34but also save it a 9% of energy consumption

2:29:36and a 9% of CO2 emission.

2:29:40Also, we commonized the SNX design in Fremont.

2:29:45Right now we have a common body and common headlamp.

2:29:49We eliminated about 40 parts and reduced about 10%

2:29:54of the cycle time.

2:29:55So all these improvements really help

2:29:58us to get to a faster ramp trajectory.

2:30:04And we didn't just learn from these high volume

2:30:06existing factories.

2:30:07We also learned from the new factories.

2:30:10Most recently, our Greek Berlin factory team

2:30:14implemented a 5G private network on the shop floor,

2:30:16helping us to overcome about a 90% over cycle issues

2:30:22for a particular process in GA shop,

2:30:26we're gonna soon implement it globally.

2:30:29So right now we have an integrated global organization

2:30:36from production all the way to a thousand delivery service.

2:30:40This will help us to a strength feedback loop

2:30:43between manufacturing and the service.

2:30:45We obviously want to bring a delight for customer ownership

2:30:49experience to our car owners.

2:30:53And with this direct feedback loop,

2:30:56we'll be able to turn customer escalations and feedback

2:30:59into a quick action improvement on the shop floor immediately.

2:31:03With that effort in the past six months,

2:31:05we'll be able to reduce the time and service

2:31:08and early ownership service, also service

2:31:10a point of a week time significantly.

2:31:15So four factories, not enough.

2:31:19If we want to hit the 20 million car a year target,

2:31:23we're going to keep building new factories, new lines,

2:31:26and that will come up with new product as well.

2:31:30And with all gigafactories, now in one organization,

2:31:35we'll be able to replicate unified ideas across factories.

2:31:39Also, we will help the new factories that rev faster

2:31:44and also produce a better quality at the lower cost.

2:31:48Now I'm going to hand over to Ju, a talk about a cell

2:31:50manufacturing.

2:31:53Thanks, Tom.

2:31:54Yeah, we're not just talking about new vehicle factories,

2:31:57but also new cell factories.

2:31:59And we're going to talk about how we follow the same model

2:32:02when we do so.

2:32:03But first, I thought I'd just provide a little bit

2:32:05of an update on cell production.

2:32:07So remember, battery day, we showed this video with the spoon

2:32:12and how we went from dry powder to film.

2:32:15Let's just say there's no spoon now.

2:32:18So many of you in this room saw this on your tour today,

2:32:22but here is our dry electrode machine here in Texas,

2:32:25one of the lines we have installed here.

2:32:27Fully automated, no spoon from powder and foil

2:32:30into coated electrode out.

2:32:33From a peak productivity perspective,

2:32:35per tool perspective, this is over 20 times

2:32:38the productivity of the tools that we showed folks

2:32:40on the tour on Kato back in battery day.

2:32:43So we've made a lot of progress on the key,

2:32:45one of the key parts of the cell manufacturing process.

2:32:49And we've also continued to focus on refining the way

2:32:53we make cells and the factories that the cells go inside.

2:32:57As you can see on this chart, from typical 2170 cylindrical

2:33:01cells to 4680 made a huge leap, which is basically

2:33:04a 5x reduction in the factory footprint and volume end

2:33:09footprint. And then from going from what we did in Fremont and our pilot line to Texas,

2:33:15we improved further and we're improving again when we go into Nevada. And what this actually

2:33:20represents is a series of actions taken by a very integrated holistic design team across

2:33:27the product design, the manufacturing design, the process design, the equipment design and

2:33:31the facility design. They all need to work together to make this happen. And you can see

2:33:36on the bottom just as an example of simplicity up investment down and scaling up, you know,

2:33:42parts.

2:33:43We've reduced the number of parts in the cell, the number of processes, significant reduction

2:33:46of number of processes, and collaborating with those five design groups, we've been able to

2:33:51result, you know, build, end up with a factor that's 10 times smaller volumetrically, which

2:33:58means it's faster to get built, it's much lower CAPEX per gigawatt hour output, and we

2:34:04We can go and scale to our objectives of 240 total terawatt hours, one terawatt hour a year

2:34:11of stationary storage and 20 million vehicles a year with the scalability that is required

2:34:17to achieve our goals.

2:34:20And we're not just looking at the self-factor itself, but also upstream materials where necessary.

2:34:26What you see here is the rendering of the 50 gigawatt hour a year, Corpus Christi Lithium

2:34:31refinery that we've already broken ground on here in Texas. The facility will start commissioning

2:34:37by the end of 2023 this year. This is a good example of something where we're basically

2:34:43talking about breaking ground and starting commissioning within 10 months and with actual

2:34:50production within 12 months, that's the target, similar to what we did in Shanghai. Again,

2:34:55the result of collaboration and internal execution

2:34:59of construction and partnership with local communities.

2:35:04This site is 30 minutes from the Corpus Christi ports

2:35:06located directly on rail.

2:35:08The process we're taking is a direct soda ash leach

2:35:12of the input material, which means there's no acid roasting.

2:35:17We don't have any of the waste products associated

2:35:19with an acid roasting step.

2:35:23We're designed to consume lithium spodramine, which

2:35:26is a very commonly traded lithium rock.

2:35:29But it is flexible to other feeds from primary and secondary

2:35:31sources.

2:35:32And similarly, we're working on our cathode facility here

2:35:34in Texas.

2:35:35Maybe you've seen it as you're driving around,

2:35:37which is a 60 gigawatt hour a year cathode facility

2:35:40behind the main building here.

2:35:4210 month build time there as well.

2:35:45The equipment's being installed for the first line as we speak,

2:35:47and we'll be commissioning starting next quarter.

2:35:50So we are doing this where we need to.

2:35:53Our plan is not to do it always.

2:35:54There are lots of competent companies out there.

2:35:57But we are also trying to sort of accelerate

2:35:59the pace of the industry by trying some new things that

2:36:01are a little bit more scalable and de-risking

2:36:05certain innovations that improve productivity

2:36:07in terms of capx per gigawatt hour and things like that.

2:36:11And with that, I think that sums up

2:36:13our efforts on the factory side.

2:36:16Mike, you want to come up and we'll talk about energy?

2:36:18All right, all right, take it away.

2:36:29All right, here you can.

2:36:31Hello, everyone.

2:36:32My name is Mike Snyder.

2:36:34I lead our megapack organization.

2:36:36I've been with Tesla for almost nine years.

2:36:39I'm excited to talk to you today about the megapack product

2:36:41and the business and some of the exciting things

2:36:43we have ahead of us.

2:36:46From the beginning with this business,

2:36:48we have always focused on building successful projects

2:36:52for our customers, and not just the batteries in the box.

2:36:56We've built a hardware and a software platform

2:36:59that is able to adapt to environments all over the world,

2:37:03and that is able to scale from small island projects

2:37:06to large gigawatt hour scale batteries.

2:37:10We've invested our time and our talent

2:37:11to best understand every aspect, every step in the process,

2:37:16and every risk of a project, to ultimately provide

2:37:20a solution that is thoughtful, intuitive,

2:37:24provides great value to our customers,

2:37:26and is as plug and play as possible.

2:37:30And yeah, just looking at these projects,

2:37:31it's incredible to see them all come together in one montage.

2:37:35Seeing the future.

2:37:36Yeah, we've compressed.

2:37:38It's 10 years of building these projects,

2:37:40and it's incredible to see the impact

2:37:42and to see what we have in front of us.

2:37:43Absolutely.

2:37:47We're on our sixth generation of our industrial product,

2:37:49which is the Megapack XL that we're

2:37:51building out of Leithrip, California.

2:37:54We've deployed over 16 gigawatt hours

2:37:56of industrial and residential products across 50 countries.

2:37:59And really, Megapack is a market leader.

2:38:01It's best in class in efficiency, reliability, energy

2:38:05density, and easiest to install, lowest cost to install.

2:38:10And we call it Excel because it is actually

2:38:12the largest, heaviest object that you

2:38:15can transfer around the roads of the world without having

2:38:19to shut them down and get crazy permits and things like this.

2:38:22So it is extra large.

2:38:24There is not going to be an XXL. Let's just say it that way.

2:38:27There's been incredible demand for the project or for the product.

2:38:322023 is going to be a great year. We have gigawatt hour scale projects being built.

2:38:36The Lathrop factory continues to scale and ramp up and there's new products on the near-term roadmap.

2:38:41So a lot to look forward to.

2:38:43How did we get here?

2:38:45Number one, a maniacal focus on all aspects of delivering stationary storage.

2:38:50What do we mean by that? What are some examples?

2:38:53As you see here, the megapack enclosure, for one example, we designed the wireways directly

2:38:58into the base of the enclosure.

2:39:00So typically you need to install a bunch of conduits and cables under the ground.

2:39:03Now you can lift them up and install them directly into the base, increasing the speed

2:39:08of deployment and lowering the cost.

2:39:10Next, we see the batteries and the power electronics.

2:39:13And the power electronics is really the beauty and elegance of the megapack.

2:39:18So this is what converts the DC power to AC power on the grid.

2:39:22The brain is built directly into the product.

2:39:25It is able to connect to any grid in the world right out of the box.

2:39:30It allows it to all the megapacks on a given site to work as one unit.

2:39:36And really, it's what makes it what it is.

2:39:38I think it's the most incredible feature of the megapack.

2:39:41As you build out the full site, what we end up with is the most energy dense solution on

2:39:45the market, upwards of 300 megawatt hours per acre.

2:39:49And just as a frame of reference, this solution is two times more power dense than a typical

2:39:54gas speaker plant.

2:39:56So this is the future.

2:39:58This is where we're going.

2:40:00This is the product that retires the fossil fuels.

2:40:03Yep.

2:40:04One power plant at a time.

2:40:05We talk about power electronics a little bit more.

2:40:08Tesla is a leader in power electronics.

2:40:10We've deployed over 1.4 terawatts across energy storage and vehicle.

2:40:16And we deliver more power electronics than the solar and the wind industry combined on

2:40:19a peranem basis.

2:40:21And power electronics, it's hard to overstate how impactful they are.

2:40:27They are really the glue in the sustainable energy economy between generation, storage

2:40:32and the end use.

2:40:35Those power electronics devices are switching thousands of times per second, hundreds of

2:40:38thousands of times per second, to efficiently react to whatever is needed, either in the

2:40:44car in the grid. And because they are so like sort of software driven at their core,

2:40:49they can provide functionality that hasn't been available to the grid in the past. And

2:40:54it's one of the reasons why renewables and stores together are such a great solution.

2:40:58Yeah. And one reason that we focus on power electronics and control so much is because

2:41:04of the impact that it has on the projects directly. So just as an example here on the

2:41:10left, we have a firmware feature that we call virtual machine mode.

2:41:15And what virtual machine mode is, is it contributes to grid stability,

2:41:20like a car shock absorber, that dampens oscillations or vibrations

2:41:24and keeps the ride smooth.

2:41:26So you can imagine if you don't dampen those vibrations,

2:41:29the vehicle could lose control or someone could get hurt.

2:41:33On the grid, you could have a blackout.

2:41:36We have one grid operator that's utilizing virtual machine mode

2:41:39and said they will not operate their grid

2:41:41at 100% renewables unless they have this feature,

2:41:43unless they have virtual machine mode working.

2:41:45Yeah, and what?

2:41:46They tried and it didn't work very well.

2:41:47And virtual machine mode, it's like synthetic inertia.

2:41:50You turn the battery power plant into,

2:41:53like through software behaving like it's

2:41:55a giant spinning machine, literally.

2:41:57And that inertia stabilizes the grid.

2:42:02But you don't need a giant spinning machine.

2:42:05You don't need like a huge fossil power plant

2:42:06or a giant hydro turbine.

2:42:07You just have the battery do it.

2:42:09And you can program it to whatever you need at that part

2:42:11of the grid, or even have it be dynamically changing

2:42:13as the grid conditions change.

2:42:14Right.

2:42:15And that's why we focus so much on it.

2:42:17We believe it's the future of how batteries are going

2:42:18to add value.

2:42:19It's going to be more about adding power stability,

2:42:21just as it is about energy.

2:42:22Yep.

2:42:23Yep.

2:42:24Auto Bitter is another software feature I wanted to mention.

2:42:26Auto Bitter is an autonomous energy trading platform

2:42:29that in its most basic sense is buying energy low and selling

2:42:33it high, and the owner can net the difference of that.

2:42:36But operating battery storage is actually very complex.

2:42:41Upwards, needing upwards of hundreds or even thousands

2:42:44of decisions every five minutes,

2:42:45it's much more complex than a PV solar plant, or a wind plant,

2:42:49or a thermal battery, or a thermal generation plant.

2:42:54And it's mostly because of the versatility and the ability

2:42:56of the battery to provide so much value in different ways

2:42:59that these decisions need to be made in real time

2:43:01in order to optimize and get the most value out

2:43:03of the batteries that can.

2:43:05So there wasn't a solution that existed in the market,

2:43:07so we built one ourselves.

2:43:09In the market that we've deployed them in,

2:43:11in Australia, in Texas, in the UK,

2:43:13it has proven to be a market leader.

2:43:15We're continuing to invest in auto-bitters

2:43:17as we expand to new markets.

2:43:21Second point of how we got here,

2:43:23a relentless focus on speed of execution,

2:43:25and there's two points to be made here.

2:43:27First one is in building factories.

2:43:30What you're seeing here is a time lapse

2:43:31of the mega-factory that was built in Lathrop.

2:43:35We took it from a JCPenney distribution center

2:43:37to a world-class manufacturing facility in less than 12 months, which is incredible.

2:43:43And really, the way that we did this is similar to what my colleagues have described.

2:43:47It's leveraging the vertical integration of Tesla.

2:43:50It's getting the vehicle manufacturing team in the same room as construction and engineering

2:43:54and making decisions quickly with all the stakeholders and decision makers.

2:43:59We're using the lessons learned from Lathrop as we speak, as we plan for our next factories.

2:44:06Second, it's about installing projects faster.

2:44:09I've mentioned this already, but plug and play is kind of at the core of what we're going

2:44:12for.

2:44:13Over the past four years, we've increased the installation speed by 4X, and we've reduced

2:44:20the total labor involved across both construction and manufacturing by 3X.

2:44:27We think that this is key to unlocking our ability to scale and our customers' ability

2:44:31to scale with us.

2:44:32We need to be laser focused on reducing that time from when the megapack leaves the factory

2:44:37to when it is operational on the grid.

2:44:42And it's not just about centralized storage.

2:44:46It's also about distributed storage,

2:44:47the future roadmap of, well, at least here at Tesla.

2:44:51And maybe some of you know about this sort of retail plan

2:44:55we have here in Texas, and I just

2:44:57wanted to talk a little bit about how that all fits together.

2:45:00So Tesla Electric really unlocks the full value

2:45:03of distributed energy and storage products.

2:45:06So it enables our customers to become their own utility.

2:45:10The data on the page is from our South Australia

2:45:13virtual power plant, another Tesla electric setup

2:45:17that we have.

2:45:19Over 2022, it's 5,000 customers that we have the data for.

2:45:23You can get a sense for what's happening.

2:45:25So just if they were being provided default utility service

2:45:28and you looked at the cost of serving them

2:45:30that electricity, on average $140 a month.

2:45:33If they all had solar and power wall,

2:45:34but the solar and power wall wasn't interacting

2:45:36with the grid, wasn't participating in the energy market,

2:45:38The cost of serving that customer would go from what would have

2:45:41go from $140 a month to $70 a month.

2:45:44If you if Tesla electric is operating those assets in an

2:45:48intelligent way to benefit the grid, Tesla electric being

2:45:51basically a software that we developed out of auto better

2:45:54for the purposes of of these distributed energy resources,

2:45:58we can actually pay the customers to bring their their

2:46:02their energy services to the grid.

2:46:04That's what happened in Australia last year.

2:46:06Australia is a little bit of a special case.

2:46:09They represent the future, South Australia in particular.

2:46:11Solar and wind supplied 70% of South Australia's energy in 2022.

2:46:16That compares to 30% in Texas and 35% in California.

2:46:20But this is an indication of where this is all headed.

2:46:24Of both centralized and distributed storage resources,

2:46:29providing the key to unlocking fully renewable grids.

2:46:34And Tesla Electric is the retail plan

2:46:36that we're using to bring that to bear for our customers that have our products.

2:46:43And, you know, this is how we're rolling it out.

2:46:47First, as it exists, it's available in Texas for people who have powerwalls in their homes.

2:46:54But we need to extend beyond that.

2:46:56Over a billion people live in markets with competitive retail electricity.

2:47:00And there's over two trillion in annual energy spend in these markets.

2:47:04This is a huge opportunity.

2:47:06And our intention is to bring this to market by market.

2:47:10And the same way that we've approached Tesla insurance,

2:47:13where we can bring value to our customers

2:47:14to reduce their total cost of ownership of our products.

2:47:18So actually what we're going to do next,

2:47:21and this is pretty exciting here in Texas by this summer,

2:47:24we're going to offer a retail electricity plan to people

2:47:27who have our cars where they can have unlimited overnight home

2:47:30charging for $30 a month.

2:47:33This is part of reducing the total cost of ownership

2:47:35of our vehicles.

2:47:37And the reason why we can do this is because Texas

2:47:39has a ton of wind.

2:47:40And in Texas, the wind blows at night.

2:47:42So actually serving these customers with electricity

2:47:45at night for their cars is the best thing to do for everyone.

2:47:49And so this is a way to incentivize

2:47:50people to charge at home at night directly

2:47:52from renewable power.

2:47:53It's part of the grand master plan

2:47:55we talked about at the beginning.

2:47:57So we're very excited about this.

2:48:00And we do see this similar to Tesla insurance, as I said,

2:48:02as further reducing the total cost of ownership

2:48:05of electric vehicles.

2:48:07So putting this all together,

2:48:09we're really at the beginning of this massive ramp

2:48:12in energy storage deployment.

2:48:14And yeah.

2:48:16Yeah, reflecting back on the master plane,

2:48:17we talk about tens of terawatt hour,

2:48:19needing tens of terawatt hours of stationary storage.

2:48:21So we have our sites set on annual production rates

2:48:24of one terawatt hour,

2:48:25which is 25x our capacity at Lathrop.

2:48:28In the near term,

2:48:30we see really strong demand

2:48:31for the Megapack product over 100 gigawatt hours per year,

2:48:35in 2023 and growing by over 100 gigawatt hours per year

2:48:38over the next few years.

2:48:39So the demand is there.

2:48:42And naturally, as we continue to focus on cost and speed

2:48:44and the value of the things that we've mentioned,

2:48:46it's clear we need to build more capacity

2:48:48and we need to ramp it quickly.

2:48:50So while the challenge is big, it's also a huge opportunity

2:48:52for Tesla, it's a huge opportunity for the Megapack

2:48:55business.

2:48:56And I'm excited to see the impact that it's

2:48:57going to make on our grid transition.

2:48:59As am I.

2:49:00Yeah.

2:49:00All right.

2:49:01All right, thank you.

2:49:02Thank you all.

2:49:03Yes.

2:49:07All right, and,

2:49:09Lori,

2:49:10Brandon, come on up.

2:49:19Hi, I'm Brandon with the legal team.

2:49:21Thank you.

2:49:22Not just for coming but being investors.

2:49:26Now we have an active and engaged board and management team.

2:49:30We've met with you.

2:49:31And today is the culmination of that.

2:49:33We've heard you,

2:49:35which is why we're excited to share this investor day with you.

2:49:38Now you will hear more from our board at the appropriate time, but for now, I'm excited

2:49:43to introduce Lori Shelby, who's going to discuss sustainability in our employees.

2:49:47Lori?

2:49:48Thank you, Brandon.

2:49:53Hi, everyone.

2:49:54My name is Lori Shelby, and I lead the Environmental Health Safety Security and Sustainability Team.

2:50:01I've been with Tesla for five and a half years.

2:50:05So I'm going to talk about Tesla's impact.

2:50:10And of course, that starts with our awesome team members.

2:50:16So you've heard a lot today about our amazing products, our process, our software.

2:50:22You've seen lots of numbers and lots of graphs.

2:50:24You've also heard a common theme.

2:50:26And that common theme is that Tesla's team is what drives our success and makes a difference.

2:50:33Whether you're in our factories, in our sales service and delivery centers, our warehouses,

2:50:40or in the field, you can see and you can feel the collaboration that's happening.

2:50:46And you also can see the pride that everyone has in working for Tesla.

2:50:54So before I go on, I want to give a big shout out and a thank you for our Tesla team globally.

2:51:00So that was great.

2:51:09That was great.

2:51:10So we are creating a lot of meaningful work through our jobs.

2:51:16We're a huge employer and we're growing every day.

2:51:20We're at 129k.

2:51:22And more than half of our team works in vehicle manufacturing.

2:51:27And almost 60% of our team is based in the US.

2:51:33People want to join our mission.

2:51:40And we need to grow and we need to hire the best people.

2:51:45Engineers wanna work for Tesla.

2:51:48Look at number one and number two, SpaceX and Tesla.

2:51:53In 2021, we got three million applicants.

2:51:59Three million, that's incredible.

2:52:04And our culture, our culture at Tesla

2:52:06is anyone at any time can raise an idea

2:52:11or make an improvement suggestion across safety, people, cost, production, quality.

2:52:22Because we have a very simple process called take charge and take charge as really employee

2:52:28engagement.

2:52:29And as take charges go up, as you can see in this graph, injuries go down.

2:52:35And you know why?

2:52:37who do the work, the people who do the work know how to improve the work.

2:52:44Sustainability, sustainability is everything.

2:52:46It's our mission.

2:52:47You've heard about it all day, part of our master plan.

2:52:51But at Tesla, we don't just tick boxes to meet a requirement.

2:52:56We really focus on the impact that our products and our software will deliver.

2:53:03In 2021, our customers avoided 8.4 million metric tons of greenhouse gas.

2:53:14That's equivalent to taking 1.7 million ICE vehicles off the road for a full year.

2:53:23Yes.

2:53:24Yes.

2:53:28We're also focusing on doing really good work and driving sustainability into our operations.

2:53:37We have year-over-year reduction in our water use, in our waste, and we're growing and building

2:53:45a renewable energy program within our operations just like supercharging, as you heard.

2:53:52So it's not just about our products, it's the way we make our products, both phases.

2:54:00Tesla, our company, is solving both sides of the clean energy equation.

2:54:05We've got the clean generation, the clean storage.

2:54:09We also have zero emission transport.

2:54:12As this slide shows, from 2012 to 2021, our products generated more energy that was consumed

2:54:21by all Tesla vehicles and our factories.

2:54:24It's amazing.

2:54:27Yeah, I like the claps.

2:54:28Yeah, I'll take claps.

2:54:31I'll take claps.

2:54:32So, our vehicles use less emissions

2:54:37than gas powered vehicles.

2:54:39Even in the worst case scenario, 100% coal grid,

2:54:45the Model 3 has less emissions than the average ice car.

2:54:51And we're only going to get better.

2:54:54You heard about that today.

2:54:56Every product we make is more efficient.

2:54:59Every factory we build is more sustainable.

2:55:02So it's only going to get better and better.

2:55:05So as we build our zero emission future, we've been talking with our stakeholders, with

2:55:14their shareholders about how do we improve our reporting framework?

2:55:21And we are aligning with the TCDF.

2:55:25So TCFD.

2:55:28So we're aligning with the TCFD, which I know many of you have been asking us about.

2:55:34I also want you to stay tuned because there's going to be much more coming in our impact

2:55:40report.

2:55:41So thank you all.

2:55:42Up next is Zach Kirkwood.

2:55:55Thank you, Laurie.

2:55:59All right.

2:55:59So this is our last section of our prepared remarks today.

2:56:05We'll kind of take what we've discussed today

2:56:07and summarize it at the corporate level for our financials.

2:56:12Maybe before I jump into the details here,

2:56:14we've never done an investor day as a company.

2:56:16We've never brought our leadership team out

2:56:19and asked them to talk about the things

2:56:20that they're working on.

2:56:22So, you know, I feel very fortunate to be able to work

2:56:25alongside this group of people and support their teams.

2:56:27And I also just want to thank them.

2:56:29They're in a room in the back listening to this for all of the work

2:56:32that they put into today and getting the company

2:56:35to where it is today.

2:56:36So, thank you to the Tesla team.

2:56:45When we were preparing for today, folks were asking,

2:56:49well, Zach, what should we talk about?

2:56:51And really, the only guidance we gave folks is, well, talk

2:56:54about the things that you're working on and talk about it

2:56:57in the context of the master plan. And what did almost everybody talk about today in great

2:57:03detail, all of the work that they're doing to take cost out? Because in this industry,

2:57:09in this business, you survive or you die based upon the ability to manage your costs. And so

2:57:15I'd like to talk about cost as well. If we look at our longest running scale production

2:57:23product, this is the Model 3, we reached 5,000 cars per week, which was our design capacity

2:57:29in mid-2018.

2:57:32And since then, we've taken 30% of cost out of this product.

2:57:37And there's two points I want to make about this.

2:57:39The first is that cost reduction, as you have heard

2:57:42throughout the course of the day, is deeply ingrained

2:57:45in our culture.

2:57:46And I think one of the most important reasons why we are

2:57:48here today as a company.

2:57:50The second point that I want to make

2:57:52is that when we're working on cost reduction,

2:57:57it's easy just to take cost out and make our products worse.

2:58:00But we have to take cost out and improve our products

2:58:03at the same time.

2:58:04This is the hard thing to do, but it's the necessary thing

2:58:07to do to continue to move forward.

2:58:09And if you look at a version of a Model 3

2:58:11that we built today, and you compare it back

2:58:13to a Model 3 that we built in 2018,

2:58:16there's a long list of improvements

2:58:17that we put into the car while also taking cost out

2:58:20of the product.

2:58:24Cost reduction doesn't just come from one place.

2:58:27There's no silver bullet here.

2:58:29And we push the boundaries on volume

2:58:31with the Model 3 program, increasing volume 3x

2:58:34over this period of time.

2:58:36We've improved productivity.

2:58:37So in our Fremont factory, we are twice as productive now

2:58:40as we were in mid-2018.

2:58:42We've made a lot of progress on overhead efficiencies

2:58:45and product improvements as we've discussed.

2:58:47And a long list of other things, including localization,

2:58:50our factory in Shanghai.

2:58:52And as we ramp up volume and as we find efficiencies,

2:58:55we work with our suppliers to do the same.

2:58:58And that leads to material cost reductions

2:59:00that improve the affordability of our products.

2:59:04And as we look forward to our next gen vehicle,

2:59:07our target is to reduce 50% of cost,

2:59:09and we've talked about that a bit earlier today.

2:59:12From going from the Model S and X platform

2:59:15to the three and Y platform,

2:59:17we took out 50% of cost,

2:59:18so the task here is to do it again.

2:59:21And this is very important because as we improve affordability,

2:59:26the number of customers who have access

2:59:28to our products dramatically increases.

2:59:30And as we link this back to our master plan,

2:59:33it enables an exponential growth in our volume

2:59:35with linear reductions in the cost of our products.

2:59:39The second point I'll make here is that,

2:59:41and again, cost reductions don't come

2:59:44from any single one place.

2:59:46And so you can see the buckets here on the vehicle side,

2:59:48battery and powertrain,

2:59:50manufacturing cost reductions in others.

2:59:52These buckets are relatively equal in size.

2:59:54And so in order to take 50% of the cost out of the product,

2:59:57we have to go through everything.

2:59:59But more importantly than just the cost of the car upfront, when we're transitioning

3:00:10to a sustainable economy, particularly with vehicle ownership, it's really important to

3:00:15think about the lifetime cost of the products.

3:00:19And this chart here is showing what the total cost of ownership per mile is over the course

3:00:26of five years.

3:00:28And we have to think about financing costs, insurance costs, the cost of power.

3:00:33We do talked a little bit about that with the plants that we're doing in Texas, wear and

3:00:37tear and maintenance on the cars, et cetera.

3:00:40And we're already at a place today in the U.S. where a base model, three on a cost per

3:00:44mile basis, is less than a Toyota Corolla, which is the highest selling car in the world.

3:00:49And as we move towards our next-gen platform, we will continue to reduce this.

3:00:53And as we work on robo-taxi variants of this platform, this cost will come down even further.

3:00:59And so this is a product that we expect to have substantially lower cost per mile than

3:01:05the highest volume products in the world.

3:01:09The next thing on cost reduction I want to talk about is operating expenses.

3:01:12We spend a lot of time on our earnings calls talking about gross margins and product cost.

3:01:17We don't spend a lot of time talking about OPEX, but I think this is one of the really

3:01:20important parts of our story.

3:01:23So since 2018, we've reduced our non-gap OPEX, 60% as a percent of revenue, and 75% on a

3:01:30per delivered car basis.

3:01:33This is quite staggering, and I want to go into the details

3:01:36about how we've done this.

3:01:38The first component of OPEX, which is R&D,

3:01:41the single most important thing here

3:01:43is that we constrain the number of programs that we work on.

3:01:48So minimizing the number of programs

3:01:49that we're running in parallel, and the key here

3:01:52is to maximize the revenue and cash

3:01:54generated from every program that we work on.

3:01:56And also maximize the technology

3:01:58sharing between the programs.

3:02:00So we talked a little bit about how power electronics

3:02:03and our energy business is shared with power electronics

3:02:05and our vehicle business.

3:02:09If we look at the SGNA side of this,

3:02:12I think the story here is pretty dramatic.

3:02:14So we're showing here a comparison of SGNA per car

3:02:17delivered compared to the traditional auto industry,

3:02:20which is both an OEM and the dealership network.

3:02:23And our estimates are that we're 60% to 70% lower

3:02:26on a per car basis.

3:02:28And if you turn this into dollars,

3:02:29this is many thousands of dollars per car.

3:02:32This is part of the reason why on the last earnings call,

3:02:35I made the comment that we think about operating margin,

3:02:39more so than gross margin as a company,

3:02:41because the integration that we have here on the SG&A side,

3:02:45particularly with vertical integration

3:02:47into our dealership network,

3:02:49this provides efficiencies that give us a competitive advantage

3:02:53from a cost structure standpoint relative to other companies.

3:02:57The other thing that I'll note here is that

3:02:59this wasn't always the case.

3:03:01And it wasn't that long ago where these bars would be inverted.

3:03:05And if we go back a little bit in time,

3:03:07and we talk about when we launched Model 3 in the Fremont

3:03:10factory, we went from 2,000 cars per week, which was SNX,

3:03:14adding 5,000 a week to that for Model 3.

3:03:17Got us to 7,000 cars a week.

3:03:19You've heard some folks here today talk about production

3:03:21hell, but I don't think we've ever talked about as a company

3:03:24as back office operations hell, which also happened

3:03:27at the same time.

3:03:29And 7,000 cars a week, 350,000 cars a year,

3:03:33we want to go to 20 million cars a year,

3:03:35and we're struggling at 7,000.

3:03:37And we're saying to ourselves,

3:03:39we have to completely rethink the way

3:03:42that we're managing our back office operations

3:03:44if we have any hope of efficiently scaling this company.

3:03:48And so this began a process

3:03:50that we're still continuing to work on today,

3:03:52but it's something that we refer to

3:03:53as the Tesla operating system.

3:03:55So in the same way that we're vertically integrated

3:03:57into the software that manages our cars,

3:03:59we've also vertically integrated into the software

3:04:01that runs our company.

3:04:03And all of our major departments in the company

3:04:06are using, not all of them yet,

3:04:09but almost all of them are using our in-house software.

3:04:12Most recently, we removed third party recruiting software

3:04:15and we're now running our own recruiting software

3:04:17as a company.

3:04:18And this matters a lot because rather than having

3:04:21a complex web of third party systems

3:04:24that are generic solution that take a lot of effort

3:04:26customized for our particular needs, we've been able to put in place with an in-house

3:04:30applications engineering team dedicated lightweight software that does only what we

3:04:35needed to do and nothing more.

3:04:37And that team works very closely with process improvement teams that sit within the business,

3:04:43who are looking through all of the processes that we're managing as a company and following

3:04:47the exact same process that Tom mentioned in the manufacturing space to look at our processes,

3:04:52process steps that we don't need, simplify process steps that are existing, and automate

3:04:57what's remaining.

3:04:59And the results of this have been quite staggering, to be frank.

3:05:03I think it succeeded our expectations as well.

3:05:07On the left side of this chart are some examples of the efficiency improvements that we've

3:05:10seen within our SGA areas as a result of this strategy.

3:05:14The North American sales team is four times more efficient now than they were in 2018.

3:05:21What are operations team financial services?

3:05:23I mean, these levels of improvement are large, and there's none of pixels on the slide here

3:05:27to list out everything, so it's just a couple of examples.

3:05:31And then in the same way that when we look to take cost out of our products, we want to

3:05:36feature up the product or improve the product while taking cost out, the same has been true

3:05:40with the use of Tesla operating system and that integration with our process improvement

3:05:45teams within the company.

3:05:46And so we've increased performance and added capabilities while also dramatically reducing

3:05:52cost on a per car basis.

3:05:55There's a list of self-service functionality that we've put in place, which customers like

3:05:59and also simplifies our operations behind the scenes.

3:06:02We've used the Tesla operating system to expand into our captive insurance base, which we are

3:06:06fully vertically integrated into.

3:06:08We run our own in-house software.

3:06:10We have our own agents, our own claims software, and claims folks who run our insurance business.

3:06:16This is also very important because as a technology company, data is very important.

3:06:22And we use these systems and the access to the data in those systems to pull out extremely

3:06:27granular and targeted reporting that enables us to see every aspect of our business in real

3:06:32time so we can make adjustments to our operations as needed.

3:06:36And there are some times examples where we say, man, I wish we were tracking this piece

3:06:39of data.

3:06:40Well, you know, it's a team's meeting away from the applications engineering team to make a

3:06:44request and then we can start tracking that data and make those changes.

3:06:48And so the feedback loop associated with process improvement inside the company is pretty astounding.

3:06:56The last thing I'll just mention here, which is something I'm personally quite proud of,

3:06:59we've made a lot of progress in our closed processes within the finance team, getting

3:07:03our 10Q and 10K filed.

3:07:05It's amongst the fastest of the 20 largest market companies.

3:07:09So great work to the finance team here.

3:07:14So all of this work on cost reduction is extremely important because we have a lot of

3:07:18money to spend ahead of us to achieve our goals within the master plan.

3:07:24So we've mentioned over the course of the day-to-day 20 million annual vehicle production as our

3:07:28target, one terawatt hour of annual energy storage production, and then expanding cell

3:07:34production service and charging in line with the growth of those other businesses.

3:07:38And so we've estimated what we think the total cost to get there will be, and there's certainly

3:07:43error bars around these numbers as we continue to progress and

3:07:46innovate. But of this, we've spent about $28 billion of that so

3:07:50far in the history of the company. And maybe this total

3:07:55investment looks large. I actually think it's quite small

3:07:57relative to our ambitions. And if you look at our 2022

3:08:01operating cash flows, and you just say, well, let's assume some

3:08:05modest growth to that, maybe not all that much, if you're being

3:08:11conservative, but the ability to pay for this level of investment from the forecast that

3:08:17we have is very achievable for us.

3:08:23I'm often asked about capital allocation.

3:08:25I think our strategy here is very straightforward.

3:08:27I touched upon it a little bit in our last slide, but just to be explicit about it.

3:08:32Obviously, the priority here is to ensure that we're using capital to run the day-to-day

3:08:38operations of the business.

3:08:39is a working capital intensive business and this has to be managed very carefully.

3:08:45And so, you know, over the course of a quarter there can be multi-billion dollar swings

3:08:49and our cash balance up and down just depending upon the timing of when we build and deliver

3:08:53cars.

3:08:54So we obviously have to have money set aside for that.

3:08:57Downside protection is another really important area where we have money set aside for that.

3:09:01And over the last couple of years throughout the pandemic I think this just highlights

3:09:05why being ready for downside protection is important.

3:09:09Preserving daily operations is the engine that then generates cash that allows us to reinvest

3:09:15into the growth of the business.

3:09:18And I'm very proud that over the last couple of years, despite the ups and downs in the

3:09:22macro environment and interruptions to supply chains, we have never once pulled back on our

3:09:27investments and growth.

3:09:28So that remains a huge priority for us.

3:09:32We do then, after investing in growth, we look at opportunistic ways to spend our cash.

3:09:37are more financially oriented where if a return exceeds a hurdle, then we can place cash in

3:09:42that.

3:09:43This is part of our story of cost reduction in the Fremont factory.

3:09:47And then on the excess side, you know, this is where, you know, the board meets on a regular

3:09:52basis and thinking about what our cash flow projections are looking forward and whether

3:09:56it makes sense to do buybacks or dividends.

3:10:03So to wrap up the finance portion of the presentation today, a couple of key messages I just want

3:10:09I'll leave you all with here is that we use innovation

3:10:14at an intense level to drive cost down

3:10:17and improve the efficiency of the business.

3:10:20And the reason this is so important

3:10:22is it allows us to improve product affordability.

3:10:25And this is particularly the case

3:10:27when we move to our next generation platform.

3:10:30Improving affordability allows us to comfortably

3:10:33make investments that grow volume, that volume,

3:10:37generates cash that then allows us to make more investments.

3:10:41And as we integrate that going forward,

3:10:44it's our belief as a leadership team here at Tesla

3:10:47that we're going to achieve unprecedented scale

3:10:49in the manufacturing space.

3:10:53And this is what's ultimately required

3:10:55to accelerate the world's transition to sustainable energy.

3:11:01So with that, this completes the prepared portion of today.

3:11:07We're going to take a short break,

3:11:09and then we'll resume with the live Q&A.

3:11:11Thank you all very much.

3:21:29Well, let's see.

3:21:29So this is probably the most significant announcement

3:21:33of the day is that we're excited to announce

3:21:35that we're going to be building a Gigafactory in Mexico.

3:21:48Yeah, and we'll have a grand opening and groundbreaking

3:21:55and whatnot.

3:21:55But we're excited to announce that the next Tesla Gigafactory

3:22:00will be in Mexico near Monterey.

3:22:03So I'm so excited about it.

3:22:05Now, I do want to emphasize we will continue to expand production at all of our existing

3:22:14factories, so including California, Nevada, here in Texas, obviously, and Berlin, Shanghai.

3:22:22So we intend to increase production at all factories, so the Giga Mexico would be supplemental

3:22:30to the output of all the other factories.

3:22:32So this is not to be clear.

3:22:34the output from anywhere to anywhere.

3:22:37It is simply about expanding total global output.

3:22:41Yeah, so it's gonna be good.

3:22:43So, peanut gallery.

3:22:51Anyway, so I think that's, yeah,

3:22:55we're quite excited to announce that.

3:22:57In fact, I believe the governor is here.

3:23:02It's hard to see.

3:23:03Welcome.

3:23:10Secretary of State, I believe.

3:23:13So at the end of the day, we're looking forward to it

3:23:16and having a big event for the great groundbreaking

3:23:19and opening.

3:23:21So let's see, with that, we can move to, I guess, Q&A.

3:23:25We've obviously got significant bench strength here.

3:23:30We should probably hit like a bench.

3:23:32And anyway, we maybe have too many people on stage,

3:23:38but we'll try to answer questions within reason.

3:23:43This is meant to be kind of more of a long-term sort of discussion as opposed to, you know,

3:23:51what will be the production for the rest of the quarter type of thing.

3:23:56So let's try to orient our questions towards long-term value creation.

3:24:03And with that, far away.

3:24:07Rob?

3:24:12It's Rob Lash for the Wolf Research.

3:24:16Very exciting plans about the next generation vehicle and power trains and batteries.

3:24:22I was hoping you can maybe talk to us a little bit about the timeline for deploying this.

3:24:28It sounds like it's more than just a vehicle.

3:24:30This is kind of a paradigm change on how vehicles are assembled, how batteries are put

3:24:36together and everything.

3:24:38Does that also just get reconfigured into everything that you do?

3:24:44The model-wise that are being built here will be built

3:24:46very differently in the future.

3:24:47Maybe just give us some feel for what happens from here

3:24:50and what's the timeline for implementation.

3:24:54Well, I'll talk a little bit about that,

3:24:55but broadly speaking, the most profound

3:25:02architectural changes will be in future vehicles.

3:25:06Retooling a factory means bringing the factory down

3:25:08for an extended period of time,

3:25:10and that's preferred not to do that, I think.

3:25:15But there are variants in how Model Y is produced,

3:25:21so we've got variants where there's a rear casting,

3:25:25where there's a front and rear casting,

3:25:28and we have the structural battery pack.

3:25:32And then there are a number of smaller improvements that occur.

3:25:35But I think for really big changes,

3:25:39those would be future vehicles.

3:25:42Yeah, I don't know if you guys want to add.

3:25:45as far as maybe?

3:25:48Me?

3:25:49Yeah, so as far as I agree with you, 100% Elon,

3:25:52it's really easy to put innovations in new vehicles,

3:25:55but long term.

3:25:56We'll obviously bring them back.

3:25:57We've always talked about that, but we

3:25:58don't want to take our factories down.

3:26:00As far as the timeline goes, we're

3:26:02going to go as fast as we can left to right.

3:26:05As always, Elon alluded to the fact

3:26:07that Mexico will build our next-end vehicle,

3:26:10but we will also be doing that in our other plants.

3:26:13And so it's really about getting them all up and running.

3:26:16We expect that to be a huge volume product.

3:26:18And now we're going to move that quickly over the next couple

3:26:22of years.

3:26:30Let's go through all of them.

3:26:32Thanks.

3:26:33First, Enora Buena, Mexico.

3:26:35Marmos.

3:26:36That's great.

3:26:36Congratulations.

3:26:38Elon, a question on applying first principles thinking

3:26:45and innovation to an area that up to now has seemingly

3:26:48been outside of your control.

3:26:51And that is on mining and extraction

3:26:53of some of the key materials.

3:26:55I believe a couple years ago you had a patent on sodium chloride

3:26:59to extract lithium from some clays and spodumine clay

3:27:03and things of that nature.

3:27:05How does that fit into the plan of maybe bringing real innovation

3:27:09into a mining sector that could use a little, maybe

3:27:12waking up and getting those costs down?

3:27:14Because that could be a real gating factor, it seems.

3:27:17Well, we're going to address whatever we think the limiting factor is at any point in time.

3:27:24So we would like to do the least amount possible.

3:27:28So we don't want to get into the mining or refining sector.

3:27:33We will do that if we have to.

3:27:36I do think the focus really should be on refining capacity.

3:27:45We need to make just a very giant amount of anode cathode, lithium hydroxide, lithium

3:27:55carbonate.

3:27:56It's really the refining capacity that is the biggest choke point.

3:28:01Yeah.

3:28:03So that's why we're building a lithium refinery in Corpus Christi.

3:28:11In terms of the mining companies that are out there and looking at that part of the

3:28:17value chain, so we do have large suppliers of lithium right now and they are aware of

3:28:24how we're approaching the corpus refinery and the technologies we're trying there.

3:28:30The reason we're making them aware of it is because we think they're fundamentally more

3:28:34scalable and as we prove them out, we plan to share that with them.

3:28:38because as Elon says, it's not really like we want to do these things.

3:28:41We're doing that because it's not happening fast enough.

3:28:43So if we can prove that it can be done faster,

3:28:45the intention is to transfer that knowledge to our large, our current suppliers.

3:28:50And the same is actually true.

3:28:52It was actually a clay process that we were playing with and we continue to work on.

3:28:56Same is true on that.

3:28:57So we've worked with our suppliers as well on trials and we're sharing knowledge there.

3:29:03And the intention is just to help the whole world do this better.

3:29:06And ultimately, this here is getting the lithium out,

3:29:10or whatever it is out of the ore.

3:29:14And we're obviously building a cathode processing facility

3:29:19just adjacent to this building.

3:29:22So a little further down the road,

3:29:23you'll see another large construction.

3:29:25That's for cathode refining.

3:29:30But we'd like to really prefer if others did that,

3:29:34we're doing it because we have to do it because we want to.

3:29:37Yeah.

3:29:37And in that case, there just isn't really any large scale

3:29:40cathode production in the United States,

3:29:41and it needed to be done.

3:29:43And again, if we're gonna do it,

3:29:45we're gonna try to do it from a first principles perspective.

3:29:47So we have tried a bunch of new things there.

3:29:49We're confident that they will work.

3:29:51And as they prove out, again,

3:29:53we want to bring them back to our suppliers

3:29:54so that they can build new facilities more quickly

3:29:57with less investment.

3:30:02Right, let's go to Ben.

3:30:04Hi, Ben Kalifer-Barrant.

3:30:05So similar on the renewable side,

3:30:08Is there more that you can do in Tesla

3:30:13to nudge the rest of the renewable industry

3:30:15to speed up since it's such a pillar of the master plan three?

3:30:22Well, I mean, I don't know what we can do,

3:30:25but I can say that if there's like entrepreneurs out there

3:30:30that want to have like a guaranteed transfer of success,

3:30:35it would be refining lithium or anode and cathodes

3:30:40or any materials whatsoever for lithium ion cells.

3:30:45It has no brainer.

3:30:48Yep, I think we're doing everything we can.

3:30:55Yeah, on the renewable energy thing,

3:30:58the thing that we can do Tesla is

3:31:01the more we reduce the cost of storage,

3:31:03the more we reduce the cost of stationary storage,

3:31:05and the more we bring flexible load to the grid

3:31:08in the form of cars charging at the right times,

3:31:10the more valuable renewables are.

3:31:13Because in Texas right now, I wasn't joking.

3:31:16There's a lot of wind.

3:31:18And at night, wind is, there's almost too much wind,

3:31:21because they don't want to turn the nukes off or whatever.

3:31:23So bringing really low-cost storage onto the grid

3:31:30makes renewables more valuable, which ultimately

3:31:32will accelerate their deployment.

3:31:34So that's how we focus on it, I guess.

3:31:38Felipe?

3:31:48Thank you.

3:31:49It's Fiti Bushwa, Jeffries.

3:31:52I've got two questions.

3:31:53The first one, when I think about in this industry,

3:31:56everybody wants to be Tesla.

3:31:57Every comic or is trying to emulate what you're doing.

3:32:01Well done.

3:32:01The one thing they don't do is focus on having as much growth

3:32:08as possible with as few models as possible.

3:32:10So I'm just trying to understand, as you

3:32:12aim for 20 million units in 2030, how many models

3:32:17do you think you need to get there?

3:32:20And how does it fit into your drive to hyperscale?

3:32:24How do you manage this?

3:32:25And also the fact that probably consumers at some point

3:32:28don't want to see a Tesla, the same Tesla,

3:32:30every street corner.

3:32:32So I'm just trying to get your sense of that.

3:32:35And my other question is on bidirectional.

3:32:39I mean, you talked a lot about making

3:32:42a world more renewables and better usage of cars.

3:32:45I think bi-directional charging is one way of better using cars,

3:32:49but you seem to have been reluctant about doing that in the past.

3:32:52So I'm just wondering what your latest views are on the topic.

3:32:59Sure.

3:33:00On bi-directional, it wasn't like a conscious decision

3:33:08to not do it.

3:33:08It just wasn't a priority at the time,

3:33:10I think is maybe the way to think about it.

3:33:12As we continue to improve the power electronics in our vehicle,

3:33:17we've found ways to bring broader actionality

3:33:20while actually reducing cost of power electronics

3:33:22in the vehicle.

3:33:23And at all things, Tesla, the goal is usually

3:33:26to get more for less.

3:33:28And so we are in the middle of kind of like

3:33:31a power electronics retool, I would say,

3:33:34that will bring that functionality to our vote,

3:33:35all of our vehicles over the next two years, let's say.

3:33:43Yeah, I guess that's how I'll say that.

3:33:45I don't think very many people are going to use

3:33:46bidirectional charging unless you have a power wall.

3:33:49Because if you unplug your car, your house goes dark.

3:33:53And this is extremely inconvenient.

3:33:59Yeah, most of the value it comes in

3:34:00charging the car at the right time.

3:34:02It's not really about sending energy the other way.

3:34:06Yeah, I mean, if you have a power wall

3:34:09that can take the house load,

3:34:10then you can use your car as a supplementary energy source

3:34:15to the power wall.

3:34:16And then, you know, you're not going to drive

3:34:18want crazy by unplugging your car, you have in the house go off.

3:34:23So I think there's some value there as a supplemental energy source down the road where if you have

3:34:30a power wall you've not diminished the convenience of the people in the house.

3:34:40And the question of number of models?

3:34:42How many models?

3:34:45Oh, I do.

3:34:46I do.

3:34:47Not that many.

3:34:48Really, ten?

3:34:49I don't know.

3:34:50Not that many.

3:34:51money. What's happened with conventional cars is people run out of things to do. So you run

3:35:00out of things to do, they just end up reshuffling the deck and you have pretty much the same.

3:35:05How many variants of car are there on the road? It feels like hundreds. But are they

3:35:13good variants? No, mostly not. They're just variants for the sake of variants. But look

3:35:20How is, how things converged with the phone?

3:35:25There needs to be hundreds of flip phones.

3:35:27Now what do we have?

3:35:29It'll be like that.

3:35:34George?

3:35:36George, Generechus from Canichor, Genoti,

3:35:38just had a question about your plans to grow market share

3:35:42in China, and also whether or not the political tensions

3:35:47between the United States and China

3:35:48impact your long-term ambitions there.

3:35:51Thank you.

3:35:57Hey, Tom, don't sweat too hard.

3:36:00Yeah.

3:36:03Well, we're still growing our market share in China, quite strongly, actually.

3:36:12Especially, you know, earlier this year, we had a price adjustment.

3:36:17After that, we actually generated a huge demand more than we can produce, really.

3:36:23And as Elon said, as long as you offer a product with value at affordable price, you don't

3:36:35have the word about demand.

3:36:37That's basically the philosophy that we follow.

3:36:39We try everything to cut costs from supply chain, from improved efficiency in the factory,

3:36:47and to pass down that value to our customers.

3:36:49I think as long as we're continuing to do that, I'm not too concerned about the

3:36:53in the market share in China.

3:36:55In the second part, geopolic, no one else.

3:36:58We do our best.

3:37:00We actually created a lot of jobs to look community.

3:37:04And we also, at our suppliers, factories

3:37:11will create a lot of jobs as well,

3:37:13and contribute a lot to the local economy.

3:37:19I think as long as we're needed in this country,

3:37:22I don't see there much of the risk of that.

3:37:27Colin?

3:37:28Colin Rush from Oppenheimer.

3:37:30One of the things I was struck by in the presentation

3:37:32was the operational efficiency metrics

3:37:34that you guys are talking about.

3:37:36Talk a little bit about the velocity of learning cycles

3:37:39and how you guys track that and think about that

3:37:41as an organization as you work into a variety of other areas

3:37:44from an innovation perspective.

3:37:52Anyone want to try that?

3:37:54I'll say one very high level thing,

3:37:56which is you can't improve something you don't measure,

3:37:58And we're like ruthless measurers at Tesla.

3:38:02And once you start measuring things

3:38:05that contribute to operational efficiency,

3:38:08you actually have a path to doing something about it.

3:38:11So I guess the key is a good measuring stick.

3:38:17OK.

3:38:19It's something I should say with respect to demand,

3:38:23which I've said a few times over the years,

3:38:25but it's sometimes, be said again, which is,

3:38:34but overwhelmingly, the desire for people

3:38:39to own a Tesla is extremely high.

3:38:41The limiting factor is their ability to pay for a Tesla.

3:38:44Not do they want a Tesla.

3:38:49It's easy for people in this room to lose sight of that.

3:38:52If your income is far in excess of what a car costs,

3:38:56then you look at value for money,

3:38:57but you do not consider affordability.

3:39:00But for the vast majority of people,

3:39:02it is affordability driven.

3:39:04This is why we cannot simply double the price of the car.

3:39:08Or you could say and think about things in the limit,

3:39:10where if you had an infinitely desirable car,

3:39:14but it costs $10 million, it wouldn't matter

3:39:19because most people do not have that.

3:39:22So demand is very much a function of affordability,

3:39:28not desire, very important.

3:39:32One of the things we weren't sure about

3:39:33was the price elasticity of demand for Tesla.

3:39:37So as we lower the price, how much does demand increase?

3:39:44And we found that even small changes in the price

3:39:48have a big effect on demand, very big.

3:39:52So that was a good thing to learn.

3:39:57Yeah.

3:39:59And then as the autonomy question is very, very big.

3:40:06Because you could potentially have, I don't know,

3:40:08five times the utility of the asset

3:40:11that you currently have.

3:40:11So passenger car is 10 or 12 hours a week of usage.

3:40:18Plus a lot of parking expenses.

3:40:20An autonomous car could be 50 to 60 hours a week or something like that.

3:40:27And you could get rid of a lot of parking expenses.

3:40:31So if this is true then as autonomy is effectively turned on for the fleet, it probably will

3:40:45be the biggest asset value increase in history overnight.

3:40:49Thank you so much Emmanuel Rosner from Deutsche Bank. So as you start launching this next

3:41:04generation vehicle and ramping up volume, what will be your nearest term priority in

3:41:09terms of segment or vehicles focus? The slide that you showed with two vehicles on the

3:41:15wrap seemed based on form factor. One of them maybe looks like a van, another one looks

3:41:20like maybe like a smaller vehicle, like potentially a model too I guess. What is the nearest term

3:41:25focus for you in terms of ramping up the next gen vehicle and how do you make sure that

3:41:31by lowering the price point so much because the cost is going down 50% you're not cannibalizing

3:41:37demand for your existing vehicles.

3:41:39I mean, demand for our vehicles in terms of desire to own them may as well be infinite.

3:41:53It's indistinguishable from infinite at this point.

3:41:57So affordability is what matters.

3:41:59So as you make the car more affordable, we will have demand to go crazy, basically.

3:42:06The issue is how do we build the cars?

3:42:07The hard part is building the cars.

3:42:09I can't emphasize that enough.

3:42:11The hard part is building the cars and the entire supply chain that goes with the cars.

3:42:16This is a logistics challenge of extraordinary difficulty.

3:42:25All the things that have to go into the car have to scale with the car while everything

3:42:29is doing an exponential ramp. And if you must even one of those things, it doesn't matter

3:42:37why, earthquake, flood, fire, revolution, I thought I heard them all. I mean, any part

3:42:47of that supply chain gets interrupted. You're now, you have a seizure. The hard part is building

3:42:56the cars by far and the splashing that goes with it.

3:43:03I mean, do you guys want to talk about the splashing stuff?

3:43:07They might.

3:43:08You know, in the presentation I'd mentioned,

3:43:11perfection is a passing grade.

3:43:13We really need everything to happen perfectly

3:43:16and the strategy for mitigating the different risks,

3:43:18some of which were anticipated and some weren't,

3:43:21is really bespoke to the situation.

3:43:23And then really the unique subject matter expertise

3:43:27and a deep knowledge of the particular supply chain

3:43:30you're managing to come up with those strategies.

3:43:32So we've seen everything from, as Elon mentioned,

3:43:35first the tariffs that flew back and forth,

3:43:38then the ocean logistics issue, a chip shortage,

3:43:42COVID floods, there was a fire and a fab in Japan

3:43:46that knocked down.

3:43:48There was a massive COVID spike in Malaysia

3:43:51a lot of the chips at the back end of a lot of chips was down there and more.

3:43:56This was one that Elon was involved with as well and yeah, it's nerve-wracking but somehow

3:44:04it works.

3:44:06Yeah and although it's difficult I think we have been laying the foundation to be as intimate

3:44:13to all the tiers of the supply chain, building that control and you know, directing the different

3:44:20years of supply chain, that's the way we are trying to mitigate the risks that come with

3:44:25it. Again, dual sourcing, triple sourcing, having redundancy is the way we've been trying to

3:44:34mitigate for it.

3:44:36Yeah, and that point, when we think about vehicles, when you think about three Ys and

3:44:41architecture, SX as an architecture, our next generation platform is more than one segment,

3:44:46and really we're thinking about all the segments that are available that we haven't captured

3:44:49and where the market would be in designing it with our supply chain partners so that we

3:44:53can go quickly through those segments for where we need.

3:44:56But to Elon's point, if you make a car desirable and affordable, you know, oftentimes it doesn't

3:45:03necessarily matter what segment it's in because it's one that you want.

3:45:08And we've seen that with Model 3 when a lot of people thought the sedan was not going

3:45:11to be a great hit, but we sell tons of them.

3:45:14So the next generation platform is not one vehicle.

3:45:17It is multiple and it's on the segment that we will really try and focus on that affordability

3:45:23and desirability point more over than where we started.

3:45:26Yeah.

3:45:29I mean, it's like, I think, it's an old saying, like, battles are one with tactics or is the

3:45:33one with logistics.

3:45:34The logistics challenge is here, our enormous.

3:45:38And when you start like being a very significant percentage of an industry, you can't over-capacitize.

3:45:45It's not realistic.

3:45:48And some of these things like you said, well, you do a source or triple source.

3:45:52You can do that maybe for small things, but you can't do it for big things because if

3:45:58you're triple sourced and one of, it's like having a plane with three engines where if

3:46:05any of the three engines fails, you crash.

3:46:07It's like, you know, so you have to eat the overcompassitize which drives your CapEx up

3:46:15and it has idle suppliers somewhere and big warehouses,

3:46:19or you design to some overage that's reasonable.

3:46:27And then you have expedite costs because inevitably

3:46:30there's something goes wrong somewhere

3:46:31and you've got to fly things around.

3:46:34So it's really just the rate of progress

3:46:37is the rate of which we are able to scale

3:46:39a 10,000 logistics problems.

3:46:43The most significant of those is the cell production.

3:46:49And so we actually deliberately try to overdo cell production or supply to have that exceed

3:47:00what is needed in vehicles.

3:47:02Because if it goes below what's needed in vehicles, then the factories stall.

3:47:10So but then what do you do with all these extra cells?

3:47:12It's like, well, okay, so the much easier thing to scale up and down is it's power wall

3:47:18and megapack output, stationary storage.

3:47:21So we can then overcapacetize in cells and packs and scale production of stationary storage,

3:47:30which is much easier to scale than vehicle production.

3:47:34So that's strategically, I think, a good thing.

3:47:37Yeah, I mean the CAPEX IV, the megapack is tiny compared to CAPEX for vehicles.

3:47:44And also megapack demand is quasi-infinite.

3:47:48It basically, as long as we are competitive with utilities, we can sell as many megapack

3:47:54of...

3:47:55Like, yeah, quasi-infinite demand for that, really.

3:48:02For multi-terra...

3:48:03Mini-terra-wide hours.

3:48:04And we got a long way to go to get to many terawatt hours per year.

3:48:10Dan.

3:48:12Let's go to Dan.

3:48:15Thank you, Dan, David, Barclays.

3:48:18I think we know competitive dynamics differ significantly

3:48:22by region.

3:48:24Cost dynamics differ by region.

3:48:25We saw that with your China gigafactory, far superior cost

3:48:28to what you had for a month.

3:48:29I realized it was a new factory.

3:48:30But they're clearly different dynamics.

3:48:33So to what extent are the cost strategies

3:48:36that you've laid out today, do they differ by region,

3:48:39or is there more of a global one-size-fits-all approach

3:48:43on reducing cost?

3:48:53Yeah, I think we're pretty consistent under the strategy

3:48:56here, but we'll try to as much localize it as possible.

3:49:01Like in China, over 95% of our supply chain are localized.

3:49:06All the way from the first tier to the second tier

3:49:08at the third tier.

3:49:10And that give us, generally, the tremendous savings on COX.

3:49:16And we also have a very localized labor force.

3:49:23And we have access to skilled labor force in the region,

3:49:26and the Yangtze-Dotter region in Shanghai,

3:49:28I'm particularly referring to.

3:49:31And among over 30,000 of employees in China,

3:49:36We have probably less than 20 expats.

3:49:41So it's a very deep localization we

3:49:44have done there to be able to have such cost structure.

3:49:48And we try to replicate this approach elsewhere

3:49:52in different gig factories.

3:49:56Of course, you know, the supplier base are different.

3:49:58The labor market are different region by region

3:50:03and country by country.

3:50:03but which are our best to localize.

3:50:09I think that's absolutely give us the advantage to compete

3:50:15with the auto EMs around the globe.

3:50:20Another thing is we have this direct selling mode, like ZAC

3:50:24said.

3:50:24We'll save a tremendous money on the OPEC

3:50:27as well.

3:50:28We have full control over all expenditures

3:50:31across the company and we don't spend money on marketing or advertising.

3:50:40Everything we saved will become the value we can offer to the consumers.

3:50:47So on that part, we also follow a pretty consistent strategy over the years.

3:50:56I mean, certainly the manufacturing stuff that Colin Pete and

3:50:58myself talked about, whether you make it here or you make it in Europe or

3:51:02make it in Asia, it applies everywhere. We think about that from the ground up, not specific

3:51:06to any region.

3:51:07Yeah, just one thing I would add to Tom's point about localization and Tilar's point

3:51:14here. I would actually see that we're moving in a direction of more standardization in terms

3:51:19of our factories and our processes and our cost reduction approaches. So as we've been

3:51:23thinking about the next generation platform, we've been thinking about the volume that

3:51:28that we aspire to build against that,

3:51:30how many individual factories do we need to build,

3:51:33and what is the fastest possible way

3:51:36to expand that footprint around the world.

3:51:39What we've done with Model Y is each factory

3:51:41is an incremental improvement of the previous factory,

3:51:44which requires engineering hours,

3:51:46engineering spend for each factory design,

3:51:48and then you end up with factories

3:51:49that are slightly different from each other.

3:51:51And as we move towards the next generation platform,

3:51:54I think the term you use Lars is copy paste.

3:51:58So to get it right from the first time,

3:52:02certainly there will be things that we learn

3:52:04and we make adjustments to,

3:52:06but try to have as much standardization

3:52:08and commonization as possible,

3:52:10which allows us to go as quickly as possible with expansion.

3:52:15Over our car on this side.

3:52:21Thanks.

3:52:22Two questions, one for Zach, one for a shock.

3:52:24So for Zach, how do you think about a long term growth

3:52:29and operating margins for the business.

3:52:31And when we look at 150, 175 billion in CapEx,

3:52:35there remained imbalance of that,

3:52:38is that essentially the CapEx guidance through 2030.

3:52:42For Ashok, the multi-trip reconstruction that you highlighted,

3:52:46how is that different for what Mobileye does

3:52:48with the REM map that they have?

3:52:51And can you give us any color on the AP4 hardware?

3:52:59So with respect to operating margins,

3:53:02You know, we intend to continue to improve operating expenses

3:53:05as a percentage of revenue over time.

3:53:08And we're continuing to take cost out of our products

3:53:11and trying to keep gross margins at a place that's healthy

3:53:13as well.

3:53:15And so, you know, from the hardware perspective

3:53:17of the business, you know, it's our expectation

3:53:20that we'll continue to stay in a healthy place over time.

3:53:24And then there's a software portion

3:53:25that's added on top of that.

3:53:27And so Elon has commented on this many times

3:53:29about the impact that full self-driving software

3:53:31can have on the economics of the business. And in that space, you know, profitability

3:53:36operating margins would be impacted dramatically. Specifically to your point about CAPX guidance,

3:53:43the intent of that slide was not to provide specific guidance, but to just be transparent

3:53:47about internally what our rough estimates are and to provide context that we think getting

3:53:54to 20 million vehicles per year in one terawatt hour of energy storage is very feasible relative

3:53:59to our expected cash generation for the business.

3:54:02That's a number that will move as we learn.

3:54:05We may choose to vertically integrate more into things.

3:54:07We may find efficiencies elsewhere to the whole conversation.

3:54:10We were having earlier about do we do mining?

3:54:12Do we not do mining?

3:54:14But the entire point is to say that this is something

3:54:17that's entirely possible based upon our forecast.

3:54:21Regarding the multitude of preconstruction, my point I was

3:54:25trying to make was that we want to auto-label most of the data

3:54:29and we can collect data from the fleet

3:54:31and then reconstruct local areas

3:54:32that can act as supervision for those slips.

3:54:35Since we want to build a scalable self-driving system,

3:54:38we don't want to rely on HD maps or anything,

3:54:40even though we could trivially build something

3:54:42using the same technology.

3:54:44But this is mostly for auto labeling.

3:54:46And this provides precise 3D labels

3:54:49for all the video sequences involved in the reconstruction.

3:54:58I don't know if we touched on this much,

3:55:00But in terms of training, we have one of the biggest

3:55:04neural net training systems in the world,

3:55:05and that we expect to increase that,

3:55:09keep believe by an order of magnitude by the end of this year,

3:55:13and probably another order of magnitude

3:55:16by the end of next year, with some combination

3:55:20of NVIDIA and Dojo.

3:55:23So by much of the scale of that is quite

3:55:27going on people. It's enormous.

3:55:32Maybe the final two questions. One from Alex and one from Chris.

3:55:42Alex Potter with Piper.

3:55:45So I definitely want to ask Drew or anybody else up there for an update on

3:55:51dry battery electrode, right? So if you're

3:55:53trying to over-capacitize towards cells,

3:55:56scale a lot of this. Clearly there's a lot of moving pieces and it's a complex

3:56:00sort of orchestra with the supply chain. But a lot of it comes down to dry battery

3:56:04electrode, at least to me.

3:56:07So how are you trending?

3:56:08That was the most fascinating part of the factory tour for me.

3:56:10And looking through that window and seeing that this is clearly

3:56:13not a science project.

3:56:15But anything that you're willing to disclose yields

3:56:19progress where you are today versus where you

3:56:20thought you were going to be.

3:56:22Yeah, thanks.

3:56:31Yeah, I mean, as you saw, right?

3:56:33This is a real factory making a lot of dry electrode

3:56:37in an automated fashion.

3:56:39We've made a lot of progress.

3:56:43It's a spectrum, like we're perfectionists

3:56:47and we have clear end goals.

3:56:51And we are, every week that goes by making progress

3:56:55towards those end goals, whether it's speed of the tool

3:56:58yield of that process or the downstream process.

3:57:03We haven't stalled out yet on the rate of progress,

3:57:07either, and that's both on the Anode and the Cathode side.

3:57:10And I think the great thing about where we are with the over

3:57:15capacity that Elon mentioned is it's giving us the opportunity

3:57:18to experiment as we go, rather than just being stuck to something

3:57:24that we happen to kick off a year and a half ago and it is here.

3:57:28Something that Elon has said to the team many times is,

3:57:32it's OK to scrap equipment or money.

3:57:34It's not OK to scrap time.

3:57:36So the way we've been approaching it is probably holistically,

3:57:40what do we think is the most likely thing to succeed?

3:57:43And even actually in the factory here,

3:57:45you saw more than one anode line.

3:57:48They are actually operating two slightly different versions

3:57:50of the final process step of the powder entering the tool.

3:57:54And it's a competition, which is going

3:57:57to be more higher yield, which is going to perform better.

3:57:59And we have the luxury to be able to do that.

3:58:01And we're taking full advantage of it

3:58:03to advance the technology as quickly as possible.

3:58:06The dry electrode problem is really quite a hard problem.

3:58:10I mean, we acquired Maxwell really just

3:58:12for the dry electrode technology, but just illustrates

3:58:18what a gigantic gap there is between something working

3:58:21at small scale and at large scale.

3:58:24We've had an extremely talented team of engineers

3:58:29working on scaling the dry electrode process

3:58:32and having it be reliable, consistent,

3:58:37and we've been grinding hard literally and figuratively

3:58:41on this for quite a while.

3:58:45It seems likely that we will be able to scale it

3:58:47to volume this year.

3:58:50Yeah.

3:58:50Yeah.

3:58:51I mean, we're basically increasing the output week over week,

3:58:54like roughly 1K a week per quarter is our internal target.

3:58:59And we're tracking to that.

3:59:01I think what you saw is effectively a ton of material per hour per tool.

3:59:09It's kind of hard to rationalize what that really means.

3:59:15But it's different than, oh, if something works in a lab.

3:59:19Well, when it's tons of material per hour, there's just different kinds of problems.

3:59:24Even if you have like 0.1% escape of like fines into the enclosure that your equipment is in,

3:59:32now you have a dust problem and like that could short out some electronics.

3:59:36It's like just things like this where when you're on a lab scale, you don't even notice

3:59:40it, but when you're doing thousands of tons over the course of months, it's like, oh, a

3:59:45new failure mode we found and that's where we're at.

3:59:47But we're knocking those out though and the team is grinding through it, but progress

3:59:54every week. And the last question from Chris. Thanks for taking a question. I have a few follow-ups

4:00:03on the next gen vehicle. First, when do you think we'll get a look at it? Maybe a prototype. Second,

4:00:10are there any details that you think you can share in terms of the size, the content, the performance?

4:00:16And then third, I think you mentioned that you would produce it in other plants in addition to

4:00:20Mexico. Should we take that to mean that you can launch it at an existing plant before you're

4:00:25finished constructing the new plant in Mexico?

4:00:30I think we'll actually have to probably decline that answer.

4:00:37We will have a proper sort of product event,

4:00:40but that would be jumping the gun if we answer your questions.

4:00:43Maybe another question, if this.

4:00:46Yeah.

4:00:47Yeah.

4:00:48I don't know.

4:00:48Anyone?

4:00:51Will Daynoff?

4:00:52I have two questions.

4:00:57This has really been a very impressive afternoon.

4:00:59Thank you so much.

4:01:00Elon, I'm curious as you've doubled,

4:01:02and you'll double again, what you've learned

4:01:06about managing a larger enterprise

4:01:08and what you might have to do to manage a bigger enterprise.

4:01:13And then secondly, I'm curious on your thoughts

4:01:15on how generative AI and these rapid breakthroughs

4:01:21in AI in the last months could help you make cars less

4:01:28hard to make. Thank you.

4:01:33I don't see AI helping us make cars anytime soon. At that point, I mean, no point in any

4:01:43of us working, I think.

4:01:44Big problems.

4:01:46We'll just chill out. I mean, I'm a little worried about the AI stuff. I think it's something,

4:01:59I don't know, we should be concerned about.

4:02:01I don't know, I think we should need some kind of regulatory authority or something.

4:02:09It's overseeing AI development and just making sure that it's operating within the public

4:02:15interest and it's quite a dangerous technology.

4:02:19I may have done some things to accelerate it, which is, I don't know.

4:02:30So I mean some of the AI stuff I think is just obviously useful like what we're doing

4:02:37with self-driving which is, you know, some people think is an AGI type problem.

4:02:46I don't think it's quite an AGI problem but it certainly requires very sophisticated

4:02:53neural nets because the road system is designed for eyes and biological neural nets.

4:02:59So naturally, the analog to that is cameras with digital neural nets.

4:03:10And the thing we found is, actually look at our neural net architecture in the car.

4:03:18It's kind of insane, frankly.

4:03:22That's upon net, that's upon net, that's upon net.

4:03:24The visualization told crashes when you open it.

4:03:27It's so complicated, but no one can visualize this right now.

4:03:30Yeah, it's hard to visually, just like literally,

4:03:32it looks insane.

4:03:35So I guess something like that's happening in our brains

4:03:38while we drive around, which is pretty wild.

4:03:41So, you know, I don't know.

4:03:51Like Tesla's doing good things, and AI, I don't know,

4:03:58this one stresses me out.

4:03:59So I don't know what to think you should say about it.

4:04:06Okay, given it's seven o'clock, I think that's all the time we have.

4:04:09Thank you very much for coming.

4:04:11And thank you.

4:04:12Thanks a lot.

4:04:13Thank you.