Q1 2023 Tesla Inc Earnings Call
Speaker 1: simultaneously make significant purchases of NVIDIA GPUs and also putting a lot of effort into Dojo, which we believe has the potential for an order of magnitude improvement in the cost of training.
Speaker 1: And it also, Dojo also has the potential to become a sellable service that we would offer to other companies in the same way that Amazon Web Services offers Web Services.
Speaker 1: even though it started out as a bookstore.
Speaker 1: So, I really think that the dojo potential is very significant.
Speaker 1: In conclusion, we're taking a view that we want to keep making and selling as many cars as we can, despite this being an uncertain macro environment. This is a good time to increase our lead further and we'll continue to invest in growth as fast as possible. Once again, I'd like to...
Speaker 1: Give a huge thanks to Oltas employees worldwide for doing an incredible job again. And yeah, super appreciate it.
Speaker 2: Thank you very much and Zach has some remarks as well.
Speaker 3: Thanks Martin.
Speaker 3: I want to start by congratulating the Tesla team for record vehicle production and deliveries. And I also want to congratulate our energy storage team for record volumes as well.
Speaker 3: There's three main points I want to make. First, automotive gross margin and operating margin reduce sequentially, but as Elon mentioned, these remain at healthy levels.
Speaker 3: In particular, automotive gross margin was impacted by a few factors since our discussion on the last earnings call, which include additional action taken in the second half of the quarter to improve vehicle pricing and one-time items, most notably warranty adjustments on older SNX vehicles.
Speaker 3: as well as increase deferred revenue for certain autopilot features as we transition technologies.
Speaker 3: Progress on vehicle cost reduction continued in Q1 with meaningful improvements on logistics and the beginnings of some commodity cost reduction starting to be realized.
Speaker 3: Per unit cost for Austin and Berlin improved as well, driven by record volumes. However, these factories still provide a margin headwind and will likely continue to do so until after we reach and stabilize at our intended volumes.
Speaker 3: Note that Q1 was our third quarter in our multi-quarter plan to move to a more reasonably balanced mix of build and deliveries.
Speaker 3: As I've mentioned previously, this results in lower deliveries and production within a quarter due to a higher volume of cars in transit at the end of the quarter and has an associated impact on quarter ending free cash flows.
Speaker 3: This was particularly prevalent in Q1 for F and X as we began exporting cars for international deliveries.
Speaker 3: Second, our storage business is starting to take shape and this is exciting to see after many years of investment in focus. This business is growing as a percentage of the business's – of the company's revenue and reached its highest level yet in Q1, driven by an increasing rate of deliveries for our MegaPak products. But this whole the President's company, and what it means hopefully, is we can manufacture
Speaker 3: We are also making progress on storage profitability, generating our highest gross profit yet in the quarter.
Speaker 3: Third, I want to reiterate the philosophy by which we are operating the business this year. Our approach is to grow volumes as quickly as possible in both our vehicle and energy businesses. Our approach is to grow volumes as quickly as possible in both our vehicle and energy businesses.
Speaker 3: We plan to continue to invest heavily into our future plans, which include the Cybertruck, Next Generation Platform, in-house self-production, energy storage business, and our autonomy and AI-enabled products.
Speaker 3: And we plan to do this while keeping the business financially healthy and industry leading.
Speaker 3: To accomplish this, we need to remain focused on cost efficiency and working capital, and in particular unwinding the strategic inventory buildup left over from the pandemic. I want to conclude by thanking the Tesla team again, as well as thanking our suppliers and our customers.
Speaker 2: Thank you very much and let's go to investor questions on sait.com. The first one is what is the process to make auto pricing adjustments? What variables do you consider? How frequently do you review pricing?
Speaker 3: Do you want to take that, Yvonne, or do you want me to take it?
Speaker 1: My apologies, sorry, I was on mute. Yeah, I think this is not something that we can really talk about. It's just we do our best to evaluate the production output, macroeconomic conditions, and we make a decision. Yeah, it looks like something you'd like to add, Jack.
Speaker 3: I think that's right. I mean, as a team we review where we stand globally on a weekly basis and certainly can't get into the details of the reasons why certain decisions are made, but it is something that's very actively managed by a subset of the leadership team.
Speaker 1: Thank you. The second question is do you still believe Tesla energy will be bigger than auto and when will you provide more formal guidance on megapack and overall Tesla energy? Yeah, I should just clarify like bigger than auto from the standpoint of like total gigawatt hours deployed. So it's possible automotive revenue may be higher but gigawatt hours I think will be higher.
Speaker 1: probably higher with stationary storage. If you just look at what's needed to transition the wealth to a sustainable energy economy, there is more stationary energy storage needed than there is mobile energy storage. So,
Speaker 1: And we are seeing growth of our station storage well in excess of automotive. So that is in line with expectations.
Speaker 3: Yeah, and on the guidance part of the question, and maybe Martin, we can combine this with the next question, which is on guidance for margins. Just have a single comment there.
Speaker 3: You know, I think we are – we will get to the point where we as a company provide guidance on the storage business. I say storage with a combination of both the mega pack business and the powerwall business. Relative to total revenues of the company, it's still fairly small.
Speaker 3: And the business has a lot of volatility currently, both in terms of volumes as well as financials, just given the small volumes and kind of diversification of the customer pool there. But as this business grows and snooze out, I don't think we're that far away from it.
Speaker 3: I think including these volumes on our day two production and deliveries release is something that we'll start doing and then we can talk more formally as a business about our expectations over the coming year. I think it'll be a few more quarters before we take that.
Speaker 3: Thank you. The next question, as you said, was already answered, so let's go to the battery question. Just one other thing I wanted to mention on margin. While we're not providing specific guidance there, just to set expectations of where we think this business will go in terms of margins.
Speaker 3: you know, probably generally in the ballpark of what we've seen historically on the vehicle business. We generally look to mid 20% gross margins for any program that we launch. And so we're not there yet on this business, but that's what we're working towards.
Speaker 1: We're hopeful to get there later this year, but that's not a promise. That's an aspiration.
Speaker 2: Thank you. The next question is how well are 4680 cells meeting the expectations described on the battery day? How long will it be until the cells meet those goals?
Speaker 3: Yeah, so on Battery Day we established a cost-down road map through 2026 across five areas of effort. There was the cell design we discussed, anode and caplet materials, the structural pack concept and the cell factory itself. We've been making progress across all these aspects since then. For the cell factory, the…
Speaker 3: The Texas 4680 factor we are partway through building and commissioning and installing and operating will be 70 percent lower capex per gigawatt hour than typical cell factories when fully ramped in line with what we described on battery day. And we're continuing to further pursue densification and investment reduction opportunities in future factory build ups like in Nevada.
Speaker 3: On the cell design, we're in production with not only the first generation tabular cell we unveiled on Battery Day, but a second, more manufacturable version of Texas Today.
Speaker 3: On the cathode material side, we have a number of activities underway per the Battery Day Roadmap. For lithium, our Corpus Christi lithium refinery breaks ground this May. Our goal is to start commissioning portions of the facility before the end of the year. The refinery uses the sulfate-free spodumene refining process with reduced process costs, passive, acoustic reagents, lower embodied energy.
Speaker 3: and actually produces a beneficial byproduct that can be reversed in purpose in construction materials. We discussed all of these concepts on battery day. Same with cathode precursor, we successfully demonstrated a lower process cost, zero waste water precursor process that we described on battery day at both lab and pilot scale and are in the detailed design phase for incorporating this technology into the front end of our Austin Catholic facility.
Speaker 3: On cathode production, we are 50% equipment and 75% utilities installed at our new cathode building in Austin with our goal to begin dry and wet commissioning this quarter and next quarter with the target to produce first material before the end of the year.
Speaker 3: Structural PAC, we saw big improvements with PAC manufacturing with the 4680 cell and the Structural PAC concept, 50 percent lower CAPEX and 66 percent smaller factory for the same output in gigawatt hours per year. You know, we do believe structural as a concept is a good one. It's simpler. We'll continue to structure it. We'll continue to structure it.
Speaker 3: Texas production increased 50% quarter over quarter, through yields increased 12%, and Cato peak rate increased by 20%, and through yields improved by 20%. All together the team accomplished a 25% reduction in COGS over the quarter, and we are on track to achieve steady state cost targets over the next 12 months.
Speaker 3: And going forward for the rest of the year, the priority one is yielding cost for the 4680 program as we steadily ramp production ahead of Cybertruck next year.
Speaker 2: Thank you very much. The next question is, what do you anticipate 2023 automotive gross margins x credits will be at the company's current pricing levels?
Speaker 4: I can start off on this one.
Speaker 4: You know, this is a difficult environment to make a projection like this. You know, there's a lot of macro uncertainty. There's also headwinds and tailwinds. You know, this is a difficult environment to make a projection like this.
Speaker 4: This is basically a question I think that's asking about, however you point out where costs will go.
Speaker 4: And within cost, there's a set of costs in which we do control, a set of costs in which we're kind of subject to what's going on in the macro world.
Speaker 4: within the bucket of things we control.
Speaker 4: Most of the cost down that we're working on is around ramping our Austin factory, stabilizing that.
Speaker 4: And then doing the cost optimization work once we get to our intended volumes there. And a part of the cost journey in the Austin factory is, as Drew mentioned, the 4680 cell, which is an input into our Austin COGS. And so, as the 4680 program improves over the course of the year on cost, as Drew mentioned, the focus was to improve the cost of payment process the end-to-end business.
Speaker 4: and then the non-fill portion of the factory improves, we see a pretty good trajectory in the Austin facility.
Speaker 4: A similar story exists in the Berlin factory. It does not have 4680 as an input, but for that factory, the journey to complete localization is still ongoing. And so over the course of this year as volume increases, the volume increases, the volume
Speaker 4: more localization occurs, we do see a good path to cost reduction in the Berlin factory as well.
Speaker 4: In existing factories too, we talk about this on every call so I don't need to rehash it, but the expectation is that every existing factory improves all of their key metrics and we continue to see the progress there.
Speaker 4: You know, there's also a handful of other costs in which we have influence, but the philosophy here is that progressively going across every cost bucket that we can.
Speaker 4: Within the world that we don't control, the two major costs there being logistics, which fortunately is moving in our favor. And I think our supply chain team has done a great job both on logistics optimization and taking advantage of reduced spot rates where they can. So thank you to our supply chain team. And then there's the commodities world.
Speaker 4: which has been a huge page point in our cost structure over the last say two years or so. And we're still kind of at the maximum of paying for commodities in our cost structure.
Speaker 4: kind of max down in the second half of last year, we did start to see in Q1 a little bit of improvement. We think there will be a little bit more improvement in Q2. But, um, it's worth mentioning that the price of lithium has dropped significantly.
Speaker 4: And that's the piece that we expect to see more impact from in Q2. And generally as a company we do expect commodity prices to come down and have a more meaningful impact in the second half of the year.
Speaker 2: So this is our approach, how that nets out, and just a lot of risk and we'll have to see how the year progresses. Thank you. The next question is, how has global order intake tracked since the most recent round of price cuts? That Marcio Armori got imagining
Speaker 1: I think the overall thing we can say is that orders are in excess of production.
Speaker 2: Thank you. And maybe the last question from investors. Can you give updated specs and pricing for Cybertruck and any new features that will make it to production?
Speaker 1: Well, I think we'll save that for the Cybertruck handover, which will hopefully be around the end of Q3 this year. And one thing I am confident of saying is that it's an incredible product. It's a Hall of Famer, I think.
Speaker 1: And a product like this only comes along once in a long while. So people will not be disappointed at all. It's amazing.
Speaker 2: Great. Thank you very much. And let's go to analyst questions. We'll start with Alex Potter from Piper Sandler. Alex, go ahead and unmute.
Speaker 3: Can you hear me? Yep. Yes. Okay. Perfect. So first question was on Lathrop. Obviously, that's great to see the growth there. Just wondering when you think that facility might be closer to full utilization. Are you just sort of deliberately working your way up the S curve there?
Speaker 3: demand obviously isn't the limitation. So what are the steps, I guess, to unlocking full utilization there?
Speaker 3: Sure. There are some classic factory ramp aspects of what's going on in Lathrop. We actually have two phases of the CAPEX there. We phased some of the General Assembly parts of the facility. But in addition, we also have ramps with our suppliers that we are following. So both on the –
Speaker 3: on the cell side and on the power electronic side. And we will see that unlocked in the latter half of this year with those inputs. So the overall facility was phased with the second phase of traffic coming online towards the end of this year.
Speaker 3: Okay, great. And then I guess my second question is on your ability to serve other markets out of Shanghai. Obviously the facility in Berlin should be opening up your ability to, I guess, allocate more vehicles to Southeast Asia, Australia, other areas.
Speaker 3: I'm just wondering what other regions you think you're maybe not yet serving effectively. What are your timelines for addressing some of those gaps in your regional exposure? Thanks.
Speaker 1: Yes, that's a good question because there are so many parts of the world that we do not.
Speaker 1: yet served with respect to vehicles especially. So we do expect to open up new markets around the world. And while those markets are not necessarily individually
Speaker 1: They do add up to, you know, if you add up a whole bunch of markets, they do collectively sum up to something significant. So it's high time that Tesla operates cars to the rest of the world, and that is something that we intend to do.
Speaker 2: Thank you very much.
Speaker 2: Thank you very much. Let's go to the next analyst, George from Kennacort. Go ahead and unmute.
Speaker 5: Hi, thanks for taking my question. I was wondering first if you could discuss your FSD take rates and whether you've seen any significant positive or negative change there. And also, given that you've reduced the prices for your vehicles, do you think you need to do that for FSD as well? Thanks.
Speaker 1: I'll be kind of answered that the details on this do take right, but the
Speaker 1: It's a tricky pricing question because
Speaker 1: The value of a card that is autonomous is enormous. So in a way, the price right now is an option value on...
Speaker 1: on an autonomous vehicle.
Speaker 1: And that value will ultimately be very significant.
Speaker 1: For those that are using the FSC beta, you can see the improvements are really quite
Speaker 1: I mean, for those that are using the FSC beta, I think you can see the improvements are really quite dramatic.
Speaker 6: you know, there'll be a little bit of but
Speaker 1: two steps forward, one step back between releases, and for those trying the beta, but the trend is very clearly towards full self-driving, towards full autonomy. And I am, for this big time, without actually getting in
Speaker 1: I hesitate to say this, but I think we'll do it this year.
Speaker 1: I hesitate to say this, but I think we'll do it this year. That's fine. I'm pretty sure.
Speaker 5: That's what it looks like. Yeah. Thank you. Maybe on the dramatic change we've seen in EV-related commodity prices, do you think it's a reflection of any recent overcapacity in mining and refining or is that sort of a coincident indicator?
Speaker 5: on global EV demand and how do you expect those prices to kind of track over the next several quarters? Thank you.
Speaker 1: And I wish I had a crystal ball to answer your question. I don't know if we can provide a.
Speaker 6: A question I would with with.
Speaker 6: a question that would have any...
Speaker 6: value really. I think we're in uncertain times and if somebody's got a crystal bowl they can lend me I'd really like to borrow it. But you know these are these are uncertain times.
Speaker 6: My guess is that economics told me weather for about a year or so.
Speaker 6: Economics told me whether for about a year or so. and
Speaker 6: people for roughly 12 months and then So this is my guess. I'm pure speculation So many weather for about 12 months and and then provided there are not no major geopolitical wildcards that that show up That that is if things start getting sunny around spring next year
Speaker 3: The only thing I would say on the EV materials markets, they're not all super liquid and some of them, for example, like less than – like single digit percentage of the market is actually traded on the spot market.
Speaker 3: And not only are they not super liquid, there's not, like storage isn't particularly fast for all of the materials. So like small mismatches in supply and demand drive like large price swings. Not really real price swings, but just like temporarily large price swings. So it's hard to read into those price swings. I don't know, Karna, if you want to add anything. We are.
Speaker 2: is going to be another oppo to basically extend that after the decade. Uh, but quantities were procuring by the spot market becaus in place and we're just g and doing more of that. T happening is because of t of the companies that are
Speaker 2: are becoming more ambitious about finding more upstream resources and exploring locations in Africa as well as South America. So that's also helping the macro situation with pricing.
Speaker 6: But just on the lithium front, to emphasize, the choke point is much more on refining capacity than it is on mining. Lithium is actually very common throughout the world, including in the US, and really never ever... It's just a very common element on earth, is lithium.
Speaker 6: So it's much more a question of where's the refining capacity and can the refining capacity keep up. That's really what matters more than where is the lithium ore. It's everywhere basically.
Speaker 6: I think that same question also extends to refining of the cathode and to some degree refining of the anode. And this is why we've...agonizing anode in yourFuck
Speaker 6: At Tesla, we're building our lithium refinery capability, Corpus Christi and our cathode refinery outside Austin.
Speaker 6: I hope other companies do the same thing.
Speaker 6: We will have by far the most lithium refining capability and the most
Speaker 6: cathode refining capability in North America, I think probably more than everyone else combined. By a lot.
Speaker 6: So can other people please do this work? That would be great. We're begging you. We don't want to do it. Can someone please? Instead of making a picture sharing app, please refine lithium. Mining and refining. Heavy industry. Come on. It's fun. It's actually fun. Yeah, yeah. Exactly.
Speaker 6: We're here, ready to buy. Yeah, that's what I'm saying. Tesla's not doing this because we want to do it. We have a lot of fish to fry, obviously, but we're doing it because others aren't doing it and we wish others would do it.
Speaker 1: Awesome. Thank you very much. Let's go to Emmanuel Rossner from Deutsche Bank.
Speaker 1: Hey Emmanuel, can you hear me? Yes, we can. Yep.
Speaker 1: Perfect. Thank you so much for taking my question. Maybe your first question for Elon on your pricing strategy. If I understand your message, you're saying, Tesla feels it's worth maximizing the volume, increasing the size of the fleet as fast as you can, because you'll be able to monetize this over the life cycle of the vehicle. Could you be a little bit more specific around?
Speaker 7: Ways you would be able to monetize sort of like this existing fleet in the future Obviously I think autonomous seems to be
Speaker 7: a big piece of it by my understanding was that robotaxi would probably be for the next generation of vehicle, not the existing one. So I guess in which ways would you monetize it?
Speaker 6: Sorry, the robot taxi terminology can be a bit confusing because that's sort of like a generic term for our next generation vehicle. And we obviously are working on next generation vehicle. It's going to be very compelling. This is just not the time to talk about it in detail as a product. So we internally call it...
robotaxi, but really all of the vehicles that have hardware 3, which is the best majority of our fleet, we believe will achieve full autonomy. So there will be a robot, like a model 3 or model Y would be a robotaxi, a robotic taxi. So yeah, that's it.
to the best of my knowledge that we believe that the current hardware can achieve full autonomy. Understood. And then maybe a question for Zach back on the automotive gross margin. So I think I guess a few months ago, even after major price cuts, you felt pretty strongly that 20% automotive gross margin was still probably a reasonable floor.
Yeah.
You know, about half of the mess against that previous conversation last quarter
is attributed to adjustments we made in pricing in the second half of the quarter.
I guess you could argue that that lowers the floor in a sense.
We've also made pricing adjustments so far this quarter, you know, so that brings it down further.
About the other half of the myth in Q1 was attributed to things that are non-recurring.
So I mentioned these in my opening remarks. The warranty adjustment for cars that were previously produced, but not part of the pedigree of cars we're building now.
And some autopilot-related deferrals as they make some technology changes here that those deferrals should get recognized once some of the software catches up. So those two things are non-repeating. So hopefully that helps answer your question.
Yeah, I mean, there's really two macro factors that are tricky.
The biggest thing, the interest rate, if there's a very high Fed rate or interest rates are very high, every time that the Fed raises interest rates, that's equivalent to increasing the price of a car. It makes the cars less affordable because people...
are able to buy cars as a function of what they can afford on a monthly basis. So that's almost directly equivalent to a price increase is any kind of interest rate increase. Then the other factor is whenever there's uncertainty in the economy, people will generally postpone.
big new capital purchases like a new car. So it's a natural human reaction.
So, you know, if people are reading about layoffs and whatnot in the press, they're like, well, they might be worried about, they might be laid off. So then they'll be naturally a little more hesitant than they would otherwise be to buy a new car.
Now this is just the nature of the auto industry. But there will be a trans-mount of pent up demand for new cars. So that goes through cycles.
Thank you. Let's go to Ben Kalo from Baird. Ben, go ahead and unmute. Hey, guys. You know, when you talk about many fish to fry, you talked about dojo being a product that you can sell. If you want to see more of Ben's morbid tour guidesuk doing a crossfire, one of themSt
outside of Tesla, how do we rank all the things you have going on and then in the economic environment? I mean, like heat pumps and everything else you have going on versus investing in the vehicle business. Is that not the right way to look at it? I'm not sure I fully understand your question, but...
I'd look at Dojo as like...
kind of a long shot bet, but if it's a long shot bet that pays off, it'll pay off in a very, very big way.
a long shot bet, but if it's a long shot bet that pays off, it'll pay off in a very, very big way.
But potentially, you know, yeah, potentially in a very, very big way.
in the multi hundred billion dollar level. But the thing that like, you know, still put it in the long shot category, but long shot with a multi hundred billion dollar.
potential outcome.
So it's a bet worth making, but not one you can sort of say like, take it to the bank type of thing. Although these days, take it to the bank. Maybe not as secure as it used to be.
So.
And I'm sure big believes in heat pumps.
And that is on our list that over time is to do a really good heat pump for homes and commercial offices and stuff. We have the technology, it's really good, but it's still a back burner item.
Our focus is very much on vehicles, autonomy, stationary storage, basically solving sustainable energy and solving autonomy which would be from solving autonomy.
If we're able to have a fleet of several million vehicles that with a software update can be potentially what several times their original value that's
If that happens, that will be the, and I think it will happen, that will be the biggest asset value increase in history, I think.
But that will be, if that happens, that will be the, and I think it will happen, that'll be the biggest asset value increase in history, I think.
That will be the, and I think it will happen. That'll be the biggest asset value increase in history, I think. Thank you, Bob.
I saw pricing, but a lot of pundits talk about the pie and losing share or gaining share. But how do you guys look at pricing versus the EVs or the ICE vehicles? Or does that not come into the equation? Sorry to ask about pricing again, thank you.
No, it's really just like, you know, every day we get a daily real time update of how many cars were ordered yesterday, how many cars were produced yesterday. We must have a.
If there's a company that's got more real-time data.
than Tesla. I'm not sure there's any company on earth that has better real-time data than Tesla except maybe SpaceX Starlink.
Because we don't have to, you know, for the other car companies they will make the cars, send them to the dealers, then the dealers will sell the cars, and then it takes quite a long time for them to get the data back to actually figure out how many cars were sold.
Like we don't have to, you know, for the other car companies, they will make the cars, send them to the dealers, then the dealers will sell the cars and, you know, and then it takes quite a long time for them to get the data back to actually figure out how many cars were sold.
whereas we know how many cars were ordered yesterday throughout the world. So our fingers on the pulse is real time and does not have latency whereas the other
we know how many cars were ordered yesterday throughout the world. So our fingers on the pulse is real time and does not have latency, whereas the other car companies have a lot of latency in their data.
as does the government. The government has a lot of latency in the data. So we're just looking at and saying, okay, what does it take to achieve a clearing price for our vehicle production? And then we make a pricing change and we see what happens immediately.
and adjust course. So we're adjusting course, and we were thinking about it literally every day, seven days a week. Every seven days a week, I look at that email, and so does the rest of the team. And we try to make the least dumb decision that we can.
I think our decisions are pretty good. Sometimes they'll be dumb, but on average, I think.
than the rest of industry. Just to add on the question about
EV market share or ICE. This comes up a lot. I think a lot of the public debate is on this concept of EV market share. You know, we don't look at it that way. I mean we look at it as a car market, not the easy market. And actually the mission of the company requires internal combustion.
I've said this for a long time, we'll look back, I don't know, assuming civilization is still around in 20 years, we'll look back on internal combustion engine vehicles the same way we look back on external combustion engine vehicles, which like the steam engines, the steam engines and the external combustion engine vehicle. You know, there's still a few around, they're kind of quirky.
You know, kind of cool collector's items. That's how our gasoline cars will be in the future. Thank you. Let's go to Colin Rush from Oppenheimer. Colin, go ahead and unmute, please. Thanks so much, guys. Can you talk a little bit about how much of the actual cost structure is variable on these vehicles? And if you could give us a range on plus or minus the lithium cost within those contracted volumes that you're seeing.
Well, I think you have to, again, we'd really love to have a crystal ball here, but we don't have it.
depending on what timescale you're looking at, most of the car is variable.
most of the car costs are variable. So, and probably if I were to guess, I think we will see improved costs from suppliers.
You know, yeah, I think we will. That is our expectation. And we're already starting to see that. Elon, I think you mentioned before, we anticipated a crash in the lithium prices. And some of that has flowed through by way of lithium carbonate reductions into battery costs. And the same thing will happen with lithium hydroxide. The length of the supply chain matters also.
because what we're talking about is very far upstream. So by the time it, you know, makes it into the battery that's in a car, it'll be several months. But, you know, beyond just the commodity pricing, as Zach mentioned earlier, we're also very focused on other metrics that make production very efficient. So, for example, detention and demerge, air expedites, I think our air expedites are down 90%, detention and demerges down 93% from the peaks. That can be hundreds of thousands of dollars per vehicle.
So we're sort of attacking all vectors and becoming very efficient.
Okay. And then my follow-up is really around stationary storage demand on the utility scale. I mean obviously there's a gigantic queue for interconnection in the U.S. And can you talk about the volume of quotation you're seeing at this point around stationary storage for that renewables queue on a global basis and how much of that is converted into actual sales?
we have visibility into the pipelines of, you know.
a variety of different renewable energy and just pure stationary storage developers, and we also develop our own projects. And we're mostly just going, we're being selective and trying to pick the products that project that best fit our mission and our objectives.
Yeah, this – again, this is not a product call, but we'll have something – I mean, we're making improvements on many fronts, including megapacks. So I think some of those improvements will improve the speed at which you can connect the megapacks to the grid.
Thank you. The next question is from Mark Delaney from Goldman Sachs. Yes, good afternoon. Thank you for taking the question. Do you still see 2 million units as an upside case for volume this year, and is the gating factor for reaching 1.8 million or 2 million units in 2023 still supply chain, as was mentioned on your last conference call, or is it more about demand at this point?
Well, you know, if you have a crystal ball, you can lend me back to the crystal ball situation.
These are volatile times. For our production standpoint,
If things go well, we've got a shot at two million vehicles is here. But that is the upside case. And we feel comfortable with 1.8.
And we'll see how this year unfolds. That's helpful, thanks. And then the company has spoken at the investor day and then for the past conference calls about opening up it.
vehicle charging network. Can you speak to some of the feedback you've been getting from both Tesla owners and non Tesla owners and how the ramp of the charging network may progress from here. Thanks.
Drew, you want to take it? Yeah. So as you may have seen, we...
Opened our 1st before before posts in Europe and and our magic dog posts in North America and Q1. And that is indicative of the direction we're heading with. Universal compatibility for all vehicles, no matter where the chart for it is, et cetera. In all major markets, and we're going to continue to roll out those.
sort of improved offerings as we build new stations. You know, we're always balancing like our ability to serve our own customers with our ability to serve new customers when doing that. I think we've been able to balance it rather well. For example, in Europe , 50% of all of our...
of our supercharging stations are open to all EVs, and we've been able to do that without any increase in wait times at all for anybody. So we're going to continue to take a similar approach as we do this in North America and China over the coming quarters.
Okay, thank you very much. Let's go to Rod Latch from Wolf Research. Hi, everybody. I just wanted to first just follow up on your comments in your letter about leveraging your cost position as others struggle with unit economics.
and also taking into account the lifetime revenue, actually in a way that most other automakers will never see, just given your service network and supercharging and other attributes. Can you just maybe give us a sense of how…
far you'd be willing to take this. Are there brackets around the range of initial margin that you'd be comfortable with? And again, any color that you might provide on the updated range of margins that you'd expect in the auto business. I think we may have...
If they raise them, that just raises the interest cost that buyers have to pay to buy a car, so it reduces affordability and therefore reduces demand.
you know, that just raises the interest cost that buyers have to pay for to buy a car so it reduces affordability and therefore reduces demand.
So it's, but if, you know, like if we look past say this year or like could go, you know.
sometime next year, middle next year. So I think things are looking really, I think, like I said, we'll bet through if there's some, you know, major geopolitical wildcard that turns up, but the absence of that, I think.
I would be very optimistic about your middle of next year and the next year.
Just to add to Elon's comments, just two other points. What's really important for us this year in addition to just managing the day-to-day of the business, but is also investing in, as Elon mentioned, what 2024 and 2025 will look like.
And so, using the cash generated from the sale of products today and reinvesting that, this is very important for us. I think that what happens to margins over the next couple of quarters only matters in the context of what that means for our ability to reinvest into 2024 and 2025. And we have a lot of space. …
before that becomes something that we have to revisit our investment plans.
And so, you know, we're planning to keep the business healthy. But I just want to caution folks about reading too much into what happens over the near term here because we're very focused as a company on making sure that when we exit this macroeconomic situation this company is positioned in the best possible way.
Yeah, exactly. Just to elaborate on on that point, though, the revenue, the long term lifetime revenue that you're targeting from each vehicle is massive.
So if you took that to the extreme, it would seem that you'd be comfortable with a relatively low initial margin. Am I misinterpreting that or is that exactly right? That is exactly right. I mean, normally in a recession when consumers feel like they're going to be able to get
I'm kind of emphasizing the whole fundamental question of affordability. For most people, their ability to buy a car is a function of can they make monthly payment or not. Like I said, if interest rates are really high like they are right now, then in some cases people can't get a loan at all.
So, I think probably banks are pretty, not leaning forward in providing loans, I expect these days. So
And I think probably banks are pretty, not leaning forward in providing loans, I expect these days, so.
You know, so that's, but like there is quite a powerful story here when you, you know, going back to me, alluded to a moment ago, I mentioned a moment ago that
Tesla is in a uniquely strong strategic position because we're the only ones making cars that technically we could sell for zero profit.
in a uniquely strong strategic position. Because we're the only ones making cars that technically we could sell for zero profit.
for now, and then yield actually tremendous economics in the future, but through autonomy.
and then yield actually tremendous economics in the future, through autonomy. No one else can do that.
I'm not sure how many of you will appreciate the profundity of what I've just said, but it is extremely significant. Thank you. Let's go to Adam Jonas from Morgan Stanley . Adam.
that's kind of intimately well over the past six months. What can you tell Tesla stakeholders?
about how an X.com or Super App could be potentially accelerated to Tesla's business model.
an X.com or Super App could be potentially accelerated to Tesla's business model?
I don't know, I guess it could make it, tends to make it easier to buy cars. So there we are, we are transferring somewhat off topic here. Okay, all right. You know, I think there's some benefit, I think probably there's some benefit. I get it, Elon. So just as a follow up.
on manufacturing. You're a student of history, and you'll know that back in 1913, Henry Ford introduced the moving assembly line in Highland Park, Michigan. And the price of a Model T, which had already been undercutting cars around the time, fell another 70 or 80 percent, and hundreds of rival car companies went into the market.
or changing supply demand in the market, but could we catalyze some Darwinian forces in the EV market?
Well, I mean, we're not trying to say take pricing actions in order to be deliberately to deliberately undermine competitors or anything like that. We really don't think about competitors that much. We just look at, you know, do people like our cars? How can we make the product better? Can they afford our cars?
I mean, we're not trying to say take pricing actions in order to be deliberately, to deliberately undermine competitors or anything like that. We really don't think about competitors that much. We just look at, do people like our cars? How can we make the product better? Can they afford our cars? And the sort of
the things like improving service and whatnot. But like I said, we do have this unique strategic advantage that we're making a car that if autonomy pans out and we think it will, where that asset is actually will be worth a hell of a lot more in the future than it is now. So it is technically possible to sell it at zero profit, but still have the net present value of future cash flows associated with that asset would be very significant.
like improving service and whatnot. But like I said, we do have this unique strategic advantage that we're making a car that if autonomy pans out and we think it will, where that asset is actually will be worth a hell of a lot more in the future than it is now. So it is technically possible to sell it at zero profit, but still have the net present value of future cashflow associated with that asset be very significant.
and service and charging and insurance and all these other ongoing revenue streams that other companies don't have. Yeah. Certainly, we want all PBs to succeed too. We just want to say that we're not in like some malicious attacks to try to discuss everybody. Definitely not. We're like opening up superchargers. We've made our patents available for free. So it's like we're trying to be helpful here, you know.
We're not trying to destroy competitors or anything like that. We're trying to help competitors, frankly, in any way that we can. Thank you. Let's go to Dan Levi from Barclays.
We're not trying to destroy competitors or anything like that. We're trying to help competitors, frankly, in any way that we can. Thank you. Let's go to Dan Levi from Barclays. Hi.
Good evening. Thank you. First question, you're ramping supply at Austin and Berlin, so I want to understand just how critical it is to further increase volume at those plans just to get the vertical integration benefits in the face of the sort of market with some demand questions. And just broadly, should we, historically you've been.
operating at the pace at which your supply allows you to produce as opposed to gauging to demand. Should we generally expect that you're going to continue to produce at your whatever the max capacity that you're allowed within your supply constraints, regardless of what the broader economic environment is, just to continue to get that volume out there?
So that is, yes, I mean, there could be like obviously a macro shock that is so severe that people just stop buying cars for some reason. But in the absence of that, we will continue to grow output at a rapid clip.
Great, thank you. And then just on the margins associated with Austin and Berlin, you mentioned Austin and Berlin have a margin drag until you reach intended volumes. I don't know if you can disclose what those volumes are. Then maybe you could just remind us of what the margin profile of Austin and Berlin will look like versus Shanghai.
once you get the vertical integration benefits in place? Well, probably one half would be quite as good as Shanghai. Shanghai has a very efficient cost structure, obviously the lowest cost structure in the world. But we do expect to be...
make significant improvements in Austin and Berlin, and continue to make improvements in Fremont as well.
We've increased our localization efforts, so that will then drive down days on hand requirements. We've made 10 percent improvement on days on hand, so we continue that path as the localization improves. For more information, visit www.fema.gov
Thank you very much. Our final question comes from Philip from Jeffries.
Yes, good evening. Thanks for taking the question. It's slightly longer term. I completely agree with your comments that we should look at Tesla in terms of auto market share and not EV market share. But I'm just wondering, as you build up the market share globally, is there a limit to the direct selling business model as you practice it?
for the direct business model as you as you practice it today.
It seems to be working well so far.
Because we hear different feedback from customers who miss the human interaction or unhappy with the service. And I'm just wondering if you're seeing some growth pains in there that would lead you to change. You're not seeing that.
Well, I mean, there are, since we're always going to have some growing pains where, you know, times this.
And it depends on which geography we're talking about, where sometimes service is behind sales, sometimes it's ahead of sales.
Tesla is growing, I believe faster than any company in history that makes a large complex manufactured object.
You know, these are, if you're trying to match, it's always difficult to match exponentials. So, but I think it is helpful to have the feedback with service because that means we feel the pain of service and then we can adjust the design to make.
the car need less service. And I think that gives us the right incentive structure because the best service is no service. The car doesn't break. And whereas if you have say a dealer network that is reliant upon service as revenue.
then you arguably have misalignment of incentives. Where they're making money on service, but actually, the best thing for the consumer is the car doesn't need servicing.
So yeah and then that's fun if I can follow up have you have you worked out I mean for many of your traditional competitors a fair amount of profits for them comes from sending spare parts and servicing you don't have that in your in your profit structure.
And have you looked at the deficit you have compared to your peers? Yeah, actually, this one is something I could wax on about for a while because really people didn't understand that the best short-selling argument against Tesla for the longest time was the fact that Tesla does not have an existing fleet. And that the auto industry, the reasons in common...
The biggest reason is that the incumbents have a large fleet, and they're able to sell new cars at close to zero margin and then sell spare parts at a very high margin, sort of razors and blades type thing.
And so the only way to actually succeed, for a newcomer to succeed, is to have a product that is so compelling that people are willing to pay a premium over the incumbent product.
And in the absence of electrification and autonomy, I don't think a newcomer can succeed.
Thank you very much, everyone. Unfortunately, that's all the time we have for this quarter. We'll see you again in three months from now. Thank you.
New and new and MO and.
of misrepresenting their intentions. Are you worried about spies? I'm really not worried about the way it's being leaked. We'll all be able to become more enlightened because we've interacted with an AGI that will help us see the world more correctly. Like, imagine talking to the best meditation teacher in history. Microsoft has been a very, very good partner for us. So I challenge the claim that next token prediction can also pass human performance. If your base neural net is smart enough, you're just asking, like.
will make a big breakthrough in their field. There's far fewer scientists who will make multiple independent breakthroughs that define their field throughout their career. What is the difference? What distinguishes you from other researchers? Why have you been able to find multiple breakthroughs in our field? Well, thank you for the kind words.
a big breakthrough in their field. There's far fewer scientists who will make multiple independent breakthroughs that define their field throughout their career. What is the difference? What distinguishes you from other researchers? Why have you been able to make multiple breakthroughs in your field? Well, thank you for the kind words. It's hard to answer that question, I mean.
I tried really hard, I gave it everything I got. And that worked so far. I think that's all there is to it. Got it. What's the explanation for why there aren't more illicit uses of GPT? Why aren't more foreign governments using it to spread propaganda or scam grandmothers or something?
I mean, maybe they haven't really gotten to do it a lot, but it also wouldn't surprise me if some of it was going on right now. Certainly, I imagine they'd be taking some of the open source models and trying to use them for that purpose.
Like I'm sure I would expect this would be something that they would be interested in in the future. It's like technically possible they just haven't thought about it enough? Or haven't like done it at scale using their technology or maybe it's happening but they don't know it. Would you be able to track it if it was happening? I think large scale tracking is possible, yes. I mean it requires all special operations. Now there's some window in which AI is very economically valuable on the scale of airplanes let's say.
in some sense it may feel like, especially in hindsight, it may feel like there was only one year or two years because those two years were larger than the previous years. But I would say that already last year there's been a fair amount of economic value produced by AI and next year is going to be larger and larger after that.
I think that this is going to be a good multi-year chunk of time, but that's going to be true, I would say, from now until AGI pretty much. Okay, well, because I'm curious if there's a startup that's using your models, right? At some point, if you have AGI, there's only one business in the world, right? It's OpenAI. How much window do they have? Huh.
does any business have where they're actually producing something that AGI can't produce? Yeah well I mean it's the same it's the same question is asking how long until AGI. Yeah. I think it's a hard question to answer I mean I hesitate to give you a number also because there is this thing where effect where people who are optimistic people who are working on the technology tend to underestimate the time it takes to get there but I think that the way I ground myself is by thinking about a self-driving car.
In particular, there is an analogy where if you look at the Tesla and if you look at the self-driving behavior of it, it looks like it does everything. It does everything. But it's also clear that there is still a long way to go in terms of reliability. And we might be in a similar place with respect to our models where it also looks like we can do everything.
And at the same time it will be, we'll need to do some more work until we really iron out all the issues and make it really good and really reliable and robust and well behaved. By 2030, what percent of GDP is AI? Oh gosh, hard to answer that question. Very hard to answer that question. Gaming over and over again.
Like the problem is that my error bars are in log scale. So I could imagine like I could imagine like a huge percentage I could imagine that they disappointed small percentage at the same time. Okay so let's take the counterfactual where it is a small percentage. Let's say it's 20-30 and you know not that much economic value has been created by these elements. As unlikely as you think this might be, what is what would be your best explanation right now of why something like this might happen? My best explanation so
I really don't think that's a likely possibility. So that's the previous to the comment. But if I were to take the premise of your question, well, like why were things disappointing in terms of the real world impact? My answer would be reliability. If somehow it ends up being the case that you really want them to be reliable and they ended up not being reliable or if their reliability is now to be harder than we expect. I really don't think that will be the case but if I had to pick one.
If I had to pick one and you tell me like hey like why didn't things work out it would be the liability that you still have to look over the answers and double check everything that's just really puts a damper on the economic value that can be used by those systems. So they'll be technologically mature it's just a question of whether they'll be reliable enough. Yeah well in some sense not reliable means not technologically mature if you see what I mean. Yeah fair enough. What's after generative models right so before you're working on reinforcement learning is this is this basically it is this a paradigm that gets us to HCI.
I mean...
hard to be specific. So you could argue that next token prediction can only help us match human performance and maybe not surpass it. What would it take to surpass human performance? So I challenge the claim that next token prediction cannot surpass human performance. It looks like on the surface it cannot. Look from the surface if you just learn to imitate, to predict.
what people do, it means that you can only copy people. But, here is a counter argument for why it might not be quite so if your neural net is, if your base neural net is smart enough. You just ask it like, what would a person with great insight and wisdom and capability do? Maybe such person doesn't exist, but there is a pretty good chance that the neural net will be able to extrapolate.
how such a person could behave. Do you see what I mean? Yes, although where would it get that sort of insight about what that person would do if not from? From the data of regular people because like if you think about it, what does it mean to predict the next token well enough? What does it mean actually? It's actually too much. It's a deeper question than it seems. Understanding the next token well means that you understand the underlying reality.
could behave. Do you see what I mean? Yes, although where would it get that sort of insight about what that person would do if not from? From the data of regular people because like if you think about it, what does it mean to predict the next token well enough? What does it mean actually? It's actually too much. It's a deeper question than it seems. Predicting the next token well means that you understand the underlying reality that led to the creation of that token.
It's not statistics, like it is statistics but what is statistics? In order to understand those statistics, to compress them, you need to understand what is it about the world that creates those statistics. And so then you say, okay, well I have all those people, what is it about people that creates their behaviors? Well they have thoughts and their feelings and they have ideas and they do things in certain ways. All of those would be deduced.
from next token prediction. And I'd argue that this should make it possible, not indefinitely, but to a pretty decent degree to say, well, can you guess what you do if you took a person with like this characteristic and that characteristic, like such a person doesn't exist. But because you're so good at predicting the next token, you should still be able to guess what that person would do, this hypothetical, imaginary person with far greater mental ability than the rest of us.
When we're doing reinforcement learning on these models, how long before most of the data for the reinforcement learning is coming from AI's and not humans? I mean already most of the different reinforcement learning is coming from AI's. Yeah, well it's like the humans are being used to train the reward function but then the reward function in its interaction with the model is automatic and all the data that's generated during the process of reinforcement learning is created by AI.
So like if you look at the current, I would say, technique paradigm which is in getting some significant attention because of chat GPT reinforcement learning from human feedback. So there is human feedback. The human feedback is being used to train the reward function and then the reward function is being used to create the data which trains the model.
Got it. And is there any hope of just removing the human from the loop and have it improve itself and sort of alpha go away? Yeah, definitely. I mean, I feel like in some sense our hopes for like our plan like very much so. The thing you really want is for
the human teachers that tell you that teach the AI.