特斯拉 (TSLA) 2016 Q3 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Tesla Motors Incorporated third-quarter 2016 financial results Q&A conference call.

(Operator Instructions)

I would now like to introduce your host for today's conference, Mr. Jeff Evanson. Sir, you may begin.

Thank you, Chanel, and good afternoon, everyone. Welcome to Tesla's third-quarter 2016 Q&A webcast. I'm joined today by Elon Musk, JB Straubel, Jason Wheeler, and Jon McNeill.

Our Q3 results are announced in the update letter at the same link as this webcast, and during our call today we will make -- we'll discuss our business outlook and make forward-looking statements. These are based on our predictions and expectations as of today. Actual events or results could differ materially due to a number of risks and uncertainties, including those mentioned in our most recent filings with the SEC. We will start today's call with some brief comments by Elon and Jason, followed by your questions and answers.

During the Q&A please try to limit yourselves to one question and one follow up. And so if you want to log into the Q&A queue, please do so now by pressing star one. And Elon, I'll pass it over to you.

Thank you. My comments will be brief because I think it's really -- what I would have to say is captured in the earnings letter. Obviously the main thing is that we are able to have our best quarter ever, achieve full GAAP profitability, and moreover I think we are headed to have a great fourth quarter as well.

One of the criticisms I've seen out there is that perhaps Q3 was at the expense of Q4. This is not true, and we currently believe that Q4 will be profitable, excluding non-cash stock-based expenses. I think there is actually a chance that we will be -- there is a chance that will be profitable even including stock and non-cash stock-based expenses. It's just a chance. It's not a promise, but I think we've got a shot at actually being profitable, even taking stock-based expenses into account.

So it's very exciting. I think we're very proud of the Tesla team for executing so well on Q3 and going into Q4 and beyond. So it has been great. Definitely one of the best moments ever in Tesla, I think. Jason?

Thanks, Elon. Just a couple points I wanted to hit on real quickly before we jump into Q&A. One is I just want to point out the prudent financial management that we have been able to accomplish over the last several quarters. An example here is back in 2015 we were spending $400 million a quarter on CapEx.

We have averaged about $250 million a quarter in 2016. That will change as Model 3 start to ramp up in Q4, but we're focusing on make sure that every dollar we spend is in its highest and best use.

From a gross margin perspective, if you look at automotive gross margin and you exclude ZEV credit revenue, we had 140 basis point improvement quarter over quarter. Lots of different factors there. One, obviously the increase in volume helps on the labor and overhead front.

Secondly, our reliability continues to get better and better. Big change in Model X over the last 12 months, as we highlighted in the letter, and continued improvements in batteries and drive units across both vehicles. Another source of gross margin improvement is supplier sourcing and the wind down of our commitments on prototype parts for Model X.

Third point on financial management, you can see our OpEx is growing sub-linear to revenue. The operating leverage that we've been talking about through the course of the year is starting to kick in. To put some real numbers around that, GAAP revenue was up 81% quarter over quarter, 145% year over year, and yet GAAP OpEx was only up 7% quarter over quarter and 33% year over year.

The second thing I want to talk for just a couple of minutes about is what we have done to the capital structure and our sources of liquidity. As you may have read in the letter, we paid down $600 million in debt within the quarter. Most notably $422 million of conversions on our 2018 converts, derisking the balance sheet in the future.

In addition to that, we were able to sign a $300 million warehouse line, which gives us more leasing capacity at great terms. The terms on that vehicle are less than 2%.

Also we managed to get an 80% increase with our largest North America leasing partner in the quarter, and we're also on task to sign up a new leasing partner in Q4. Generally I would just like to point out that our access to capital markets and our sources of liquidity is a strong as it's ever been.

Just to highlight one element of what Jason is saying, our vehicle gross margin increased Q2 to Q3. One of the other things I've seen out there is that somehow we achieved these numbers as a result of widespread discounting; that is absolutely false.

The discounts, there were a few discounts, but they were few and far between and that has been absolutely shut down to zero. So you can see that in the fact that the vehicle profitability increased, even excluding ZEV credits from Q2 to Q3.

All right, Chanel, I think we're ready for the first question.

Colin Langan, UBS.

Great. Thanks for taking my question. It looks like a very strong free cash flow quarter, but when I look through the balance sheet there seems to be a pretty large increase in accounts payable and accrued liabilities that seems to have helped.

How should we think about that going into Q4? Does some of that unwind? Were there any changes to higher terms in the quarter, or is that just consistent with the ramp in production?

Sure. Great question; this is Jason. So there was definitely an increase in payables, and I think that will start to unwind a little bit in Q4. A lot of that is natural. If you look at production I believe it increased 37% quarter over quarter. So there is naturally going to be more parts coming into the factory.

So I think some of that is just in the course of business. The other thing that I think is worth pointing out on the cash flow statement is receivables. We had a lot of deliveries right at the end of the quarter, so we were not able to collect all of our receivables. We ended up with a fairly large receivable balance on cars that were delivered in that last 10 days or so.

Yes. Definitely also with emphasizing that, it's a first approximation you expect payables to increase by 37% if production does. And you have to net out against receivables, and when you do that I think it is not really a material situation.

And we are actively looking to increase terms with suppliers, and I think as our production has been more predictable our suppliers have been much more open to that conversation.

In fact -- thanks for making that point, Jason. I think it's worth emphasizing that for Model 3, but the Model 3 system is designed so that -- the whole manufacturing supply chain system is designed so that the faster Model 3 production grows the faster Tesla's cash balance grows. So the terms that we're getting from suppliers are significantly better, almost 60 days. As we've had to about 40 or 45 days for S and X, and the Model 3 production and logistics is way faster.

The car spends much less time in the factory, and we're working on ways to expedite delivery of the vehicles to the end customer, which we can do when we have scale. We don't have to just wait for a ship to go somewhere.

We can fill up the whole ship and just have the ship go anywhere we want. So the net effect is that instead of growth being a capital consumer, growth is a capital producer.

Got it. And so the other question I had is, your guiding to profit in Q4 without ZEV credit. It actually sounds like without even the stock comp possibly, yet production is about flat, Model X (inaudible) is going to get a little worse, OpEx guidance implies that's up sequentially. So what are the key drivers that are actually going to get you to profitability? Because I think if you take out the ZEV credit it would've probably been still a loss in this quarter.

We expect gross margin to increase, and that's a huge factor. (inaudible) We're using very few prototype parts, or low-volume selling parts, and we're not paying for crazy amounts of expediting, and there are a bunch of design improvements, design cost downs that actually are value neutral to the customer or in some cases actually it costs slightly better. And we have the P100. One of the things that the 100 kilowatt hour car [pack] was only in limited production towards the end of last quarter, and it will be a pretty significant portion of this quarter. So, yes.

One additional thing is that reliability of the cars continues to get better, so our warranty costs start decreasing as well. And that's a really -- that's a key driver for us, not only from the cost side but from the demand side where we're creating demand in market given the reliability of the vehicles.

The reliability improvement is massive.

It is. So the visits to service for Model X through the course of the year have declined 92%, which is just a fantastic result both from the manufacturing side, and the vehicle reliability teams have been working hard to achieve that, and we're going to continue to approve against that.

Chanel, let's go to the next question, please.

Brian Johnson, Barclays.

Good afternoon. I just want to go in a little bit on the regulatory credits. A couple things, first clearly with your delivery numbers the California and the other CARB states are buying more Teslas, yet last quarter you talked about the value of those plummeting and we shouldn't really expect much.

So, obviously you are generating more, but a couple questions. What's happened in the marketplace for those credits? And I know even under GAAP you don't list that as a balance sheet asset, but if we were to think about the quarterly generation of credits as well as the credits on your -- in effect in your car bank that could be monetized in the future. How would we think about this?

Unfortunately, as I've said on record before the CARB, their credit mandate is incredibly weak and needs to be fixed. And when you have a weak mandate obviously the value of those credits decline considerably.

There are some quarters where we simply cannot even find a buyer for credits, and then when we can find a buyer it's typically at $0.50 on the dollar for the ZEV credit. So -- and then obviously the ZEV credit is only applied to roughly half of our market in the US, maybe slightly above half.

It doesn't apply to Asia or Europe or Canada or Mexico or anywhere else. So it's fair, and I think CARB really should be doing more, it's unfortunate that they are not. And then I need to maybe write a longer blog piece going through this, but Tesla is sometimes criticized for relying on tax credits and that kind of thing, people really misunderstand this.

What matters is what does Tesla receive relative to its competitors, not what does Tesla receive in the absolute. But our competitors maybe worth noting maybe you would consider those to be a risk or something that is problematic for us, our competitors monetize ZEV credits at a hundred cents on the dollar.

We monetize them at $0.50 on the dollar where we can get it. That means if you have, let's say, it depends on the scenario but if you have three ZEV credits for an EV they would essentially be worth $5,000 each. So that would be $15,000. So when say, Jim or somebody sells an EV they get $15,000. But when Tesla sells an EV we get half that.

Right, they have an internal market.

It's not we who are being subsidized but our competitors. Now the interesting thing is that there is a limit to our disadvantage. Because for them the credit thing is so weak, it only goes so far. It only applies to certain states.

So what you will see our competitors do is they will limit their production and they will only sell in ZEV states, or almost entirely in ZEV states. That doesn't scale. That will take them to maybe 40,000 or 50,000 units a year best case, but not -- but we're talking about doing 500,000 units a year.

Which means at high volume we no longer suffer the disadvantage of the credit regime. This is wholly misunderstood.

Does that help, Brian?

Yes, just a quick follow up on Cap A credits. Were there any?

Say again, Brian?

Just a follow up. Were there GHG or other Cap A credits? And how did they compare to prior quotas?

Those are mouse nuts.

Okay. Thanks.

Let's go to the next question, please.

Colin Rusch, Oppenheimer.

Thanks so much. Can we just look at the shipment numbers? So a quarter ago you were guiding to roughly 80,000 vehicles a year, and now three months later we're down at 75,000. Can you just walk us through the factors that are impacting that lower shipment number, or delivery number I should say?

I think this was really gone over in last quarter's call, is that we had problems getting to rates in the first half of the year. Rate being an average of roughly 2,000 cars a week. Just a lot of things broken in our production system. I personally probably took a year off my life or more camping out at the frigging factory solving that along with a number of other members of Tesla team.

We went through bloody hell in the first half of this year. We got out of that basically around mid-June, and then the result is achieving a weekly production target of roughly 2,000 cars a week.

Okay, maybe I can take that offline. So then the second question for me is really about absorption. With nearly 40% increase in deliveries can you break out the impact in gross margin to absorption? It would seem that, that would be a meaningful number at this point.

What are you talking about?

Factory absorption.

Do you mean fixed costs versus -- ?

Yes, fixed costs on the factory and how that flows through the depreciation line.

The way to think about that is, I think Elon actually just covered it in his last answer, we had capacitized the factory and had the factory essentially produce many more cars in the first half of the year, and we fell short of that. And now we're at the rate that we had planned to be at early in the year, so our absorption is about what we'd expect it to be. And I think what you're seeing now from an absorption perspective as it's related to gross margin is a good steady-state rate.

Okay, I was just looking for a quarter-over-quarter number in terms of the contribution margin.

We typically don't break down all the different factors within gross margin.

Okay. Thanks a lot.

Ryan Brinkman, JPMorgan.

Great. Thanks for taking my question. Can you talk about the drivers of the substantially less than expected capital expenditures in the quarter and the reduction to the full-year CapEx guide? Should we think about this as being more about the push out or delay of certain activities that give rise to CapEx, or is it more that you were on schedule with those activities but doing them in a thriftier way or some sort of combination of these factors?

One of the things that we found is way better with the 3 program than X is, is that our equipment suppliers are willing to work with us on payment terms, and we've been able to backend load and in fact post production load a huge amount of the CapEx. So that just turned out a lot better than we expected, but we're not taking any action that would cause the Model 3 timeline to be extended in any way.

Okay, great.

We're still highly confident of reaching volume production in the second half of next year.

If I might chime into that a tiny bit as well, we also are continuing to improve the capital efficiency per unit of the production lines, and especially over the last few months. We've put a huge amount of engineering attention into really focusing on that problem and we're seeing results, and I think we will continue to see incremental improvements all the way from things like the battery cells all the way up to the vehicle itself.

Okay, and then the follow up to that is just in regards to the amended F4 that you filed a couple of weeks back, there was some changed language in there from Tesla is currently planning to raise additional funds by the end of the year to, now stating that you expect adequate liquidity through at least the end of year, I think it says. So what was the primary change would you say? Does it relate to this CapEx issue that we're talking about here, or to higher earnings or to another factor?

I think it covers all of the above. We've gotten really good at capital efficiency. JB, who was just speaking, has done a great job of that up at the Gigafactory in particular.

And I think we're just executing very well. We met our internal targets for Q3, so -- and you see what happened on the cash flow statement. So I think it's operational execution as well as capital efficiency.

Great. Thanks a lot.

One thing that's worth mentioning, and certainly I would take this with a grain of salt, sometimes I will say things which I think are speculation or my best guess, but they are not -- [different] from a promise. Our current financial plan does not require any capital risk for Model 3 at all.

Now that's different from saying whether we should raise capital or not to account for uncertainty to have a larger buffer and to derisk the business. And then, we also feel pretty good having examined the SolarCity financials that it looks like SolarCity will actually be at least neutral, but perhaps a cash contributor in the fourth quarter in a small way.

But again, do not take this to the bank, that's not a promise this is like -- this is what appears to be the case. Contingent upon shareholder approval, we expect SolarCity to be somewhere between neutral and a cash contributor in the fourth quarter. And yes, things are looking good.

It's not to say that there is going to be some darkness ahead, but they look really quite good right now. It seems like we probably won't want to do a capital raise even in Q1. I'm not saying we won't, but probably not. And it's overall looking quite promising.

The other thing I would just add on top of that is I would just go back to some of the comments I made at the beginning of the call about our other sources of liquidity. I mean the capital markets are open to us, and as our asset base grows our ability to monetize those assets increases.

We've got our AVL line, we've got the $300 million warehouse line. So we've got those things, and we've also been able to line up a bunch of incremental capacity on the leasing side for the rest of the quarter as well. So that's definitely a piece of it.

Very helpful. Thank you.

Emmanuel Rosner, CLSA.

Good afternoon. I have a couple of questions on your recent announcement around autonomous driving. So the first one is on hardware and then the second one on the software.

On hardware, it seems like at least from the outside where we're sitting, it seems like just recently you were indicating you will be deemphasizing the vision approach to ADAS and autonomous driving, and now it seems the latest hardware seems largely based on vision. So was curious how -- what was the thought process there?

And still within hardware, how did you acquire confidence that the hardware you're putting in cars today will still be adequate to take you all the way to full atonomy when it is only based -- I mean largely based on vision?

First of all, I would separate what Tesla says from say some supplier of ours issuing (expletive). The blog that I wrote was very clear that radar is moving from a supplemental to also a primary sensor. It's not to the exclusion of vision, but it's also a primary sensor. Vision is still the main thing.

But radar, instead of merely being like a crosscheck against vision is really when done well, and we're very confident at this point that it can be done this way, can be a primary sensor such that you can take actions based on radar information. You can also take actions based on just vision alone, much as a person might take action based on whether you hear something or you see something, but you don't need to both hear it and see it.

So we feel highly confident that the 8 camera solution with 12 ultrasonics and a forward radar and the computing power that we now have on board is capable of full autonomy at a [simply] greater than a human. There are obviously skeptics out there; well I suggest that they do not bet against us.

Okay. And then on the software side, I guess a lot of the players involved in developing autonomous solutions seem to think that a big input for autonomous driving, especially high levels of autonomy, serve a live updated map. What are -- there was not a lot of new information on the most recent announcement on this. What are Tesla's plans for this part of the solution?

You mean --?

I think we're getting into technical questions that are not really related to this quarter, so we will have to pass.

Stay tuned for product announcements as they come out.

Got it.

Okay. Thanks, Emmanuel.

John Murphy, Bank of America Merrill Lynch

Good afternoon. Just a somewhat of a redundant follow-up question here, but I really just want to make sure I get this right. As you're looking at R&D and CapEx, those are two items that as we're looking at a very significant product launch next year are running at very low levels.

I'm just curious, as you are talking about this, do you think that -- well I mean, no, they're not that low, but relative to what we would expect ahead of a product launch. Do you think that R&D at absolute levels can stay here and support the Model 3 launch and everything else you are working on?

Or will that need to go up? And then also similarly, this CapEx number of $1 billion plus in the fourth quarter really is a significant step up. Is that really just too high a number and you really are running significantly lower than this, 2.25% lower than $1.8 billion, maybe something significantly lower, and really finding a massive amount of efficiency here? I'm just really trying to understand what these levels are going to be because they are very impressive to date.

I can take the R&D piece. I imagine that R&D will continue to go up.

Not in giant ways. Moderate increases in R&D are to be expected, but not some big step change.

Exactly. On the SG&A side that is where we're really finding a lot of operating leverage. On the capital front, again, I think there's just continued opportunities for us to optimize this. There's a whole new paradigm of thinking that we're going through, and it's breaking through conventional norms such as to add a step change in capacity you have to have had a step change in capital.

That is not true. You can always optimize things. You can make things move faster. It could be more efficient, you can use floor space better. So I think it's some of this thinking which Elon has talked a lot about that is really getting baked into our capital plans.

Maybe it's seems low relative to the traditional industry, but I guess if we're comparing to what we've done in our past, and even if we just look at the --

It looks like a lot of money to us.

The S program was actually quite a lot lower R&D and lower CapEx than this. It feels like a huge amount of money.

But I mean you really are running at a run rate that is half of what you, or less than half of what you originally were talking about for the year on a run rate basis. I'm just trying to understand if that's something that is more realistic, or we should expect a real big step up in the fourth quarter.

If you go back to actually our guidance at the beginning of the -- on the OpEx side, I believe our initial guidance was 25% year over year, and we bumped that up to 30% year over year. But there is --

I'm sorry, I mean on CapEx.

On CapEx our original guidance at the beginning of the year was $1.5 billion. And then when we made the initial announcement to bring forward production of 500,000 vehicles into 2018, then we bumped it up. And I think now we're just getting smarter about that and that's why we brought that guidance back down in the letter this quarter.

We're probably just too conservative on how our capital projections. It's turned out to be we can do this with less capital than anticipated.

Just one follow up on mix. These 100 kilowatt hour models, it sounds like in some ways you may have underestimated the high end of the market, which is a good thing. As we think about that as a percent of mix going forward, do you really think there's a tremendous opportunity for that to be a material part of the mix?

Yes. Right now there are three things that are top priorities for me. Obviously, Model 3, achieving our rate at schedule in Model 3 -- rate at schedule and cost on Model 3 is top. Then it's advancing the autopilot software, autopilot to self-driving software, and then it's the 100 kilowatt hour, trying to ramp up the 100 kilowatt hour production rates.

I receive daily updates on the 100 kilowatt hour production. After this call I'm probably going to be on the 100 kilowatt hour production line, because the demand is high and we just need to satisfy that demand.

It seems like that almost might be more important as far as profitability and cash flow in the near term than the Model 3.

Definitely in the near term. (laughter) That's 100% certain.

But even over time. But okay. Thank you.

It's a very big deal. Seven days a week I get an update on the 100 kilowatt hour progress on the production ramp of that.

Thank you.

Ben Kallo, Baird.

If I can ask a question about SolarCity, one of the things when I cover SolarCity, and you bought Silevo, my initial reaction was to be negative on it. And one of the things I've worried about with the transaction if you acquired or merging is the Buffalo deal and it just being a cash cow.

And so, I was relieved when I saw Panasonic step in. Can you talk anymore about that? And then maybe in the same question I saw this -- the slide deck yesterday about how their business model is changing from lease to more cash sales or loan sales.

What do you expect going forward, and maybe that's Jason, from a cash flow basis? I know you said Q4 or they had said that, but can this be cash accretive to the business next year? Thanks. And I have one follow up.

I think, I expect SolarCity to be approximately cash neutral, all things considered, next year. Yes. It does depend on how fast we ramp up production in Buffalo. And by the way, I think previously in your question you said cash cow, I think maybe you meant to say cash vacuum.

In your questions, that's not what I'm saying it is. We do think it's important to have tight control over the production of the solar panels in order to really -- in order to have a beautiful solar roof product, we've got to be able to iterate rapidly and have them made exactly the way we want them so that you have very high efficiency sales at the lowest cost. That's our objective.

Just as we've been able to achieve that in partnership with Panasonic on the battery front; we have the best cell at the lowest price. That's a really good place to be and we're confident we can achieve that same outcome in solar. And while also creating a solar roof product that is better than a normal roof, looks better than a normal roof.

The [marker of C5 kit], as I mentioned before, if somebody has just installed a roof and the house is new, it's not going to make sense for them to go reroof the house. It makes more sense to have something that is -- solar panels that are added to the roof. But both someone that is building a house or where the roof is nearing its expiry date, then the solar roof is the right option.

So the nice thing is you don't really cannibalize one from the other, they are two separate markets. And I think you will be quite pleasantly surprised by what we debut on Friday. It's exceeded my expectation. I don't want to jump the gun on that. You should really see what we unveil on Friday. I think it's really great.

And then one more, maybe big picture and I will probably get made fun of this for asking, but stakeholders -- I've watched the work you're doing with SpaceX, and the statement you said about the reason you want to make money is for your work on interplanetary transport. How do you judge a Tesla shareholder versus a Tesla car holder? How do you delineate between where you give value versus the different stakeholders in the whole group there?

I don't really think about it like that. It's really just we want to make products that people love and then make enough money up from that to be able to develop new product, and that's it. Really it's just -- there's so few products, how many products can you buy that you really love? It's so rare.

I think if you do something like that people will buy them. They will pay a premium for something that they love, of course. Then I think it ends up being a good outcome for shareholders because the whole purpose of any company existing is to make compelling products and services. Sometimes people lose sight of why companies should even exist.

For example, would you scale back the growth of SolarCity, even though it's the greater good for the environment to be more cash flow positive, I guess is a good way to look at it.

I don't think -- we have to look at this in long term, and if SolarCity is moving lots of money then that is not good for the long term. Investors will not support such a situation. So I think there may be some intermediate slowdowns, but this is actually with an eye towards ultimately moving way faster.

Got it. Thank you.

James Albertine, Consumer Edge.

Great. Thank you and good afternoon. I wanted to ask a question if I may on battery costs, and particularly just an update on the Gigafactory and the impact of the Gigafactory on battery costs. It seems as we're getting closer to the opening that while there is some improvement going on in the background in terms of efficiencies of the battery cell production process and the trade secrets that you're working on between generation to generation of cell production, that at least 30% benefit from the Gigafactory. How should that filter into the model, of let's say over the course of the next six to eight months, between now and maybe when you start to talk more about Model 3 production? Thanks.

Well, I'm not sure we want to give a detailed glide slope on this, but we're still very confident on the progress against the milestones we talked about previously. We're still confident that we'll have the best cell cost in the world when we start production, and I think those are really the most important metrics. In long term we see ongoing opportunities to keep driving that down as we add innovation into the manufacturing process and keep increasing scale.

And just to confirm when you're still expecting to begin production on the Gigafactory itself, and then just a quick product question. As it relates to what you've done with the Model S and the 100 kilowatt battery pack, thinking about a fully loaded optimized Model 3, which obviously is a smaller vehicle, is there a potential to see range? Again, this not a price sensitive question, but at the high end of the Model 3 side, is there a potential to see range extend significantly further than what we're seeing with the S and X? Thanks.

Well maybe to your first question, we're still generally on track, as we stated, with the Gigafactory schedule and production. There's equipment being installed and being commissioned as we speak.

There is a fairly extensive process of bringing that equipment online, starting up pilot production, validating the pilot production. So that is exactly what we're in the middle of and continuing to ramp up through the end of this year. So we feel good about where that's at and we feel that we are definitely on schedule for production for Model 3.

Great. Thank you.

Adam Jones, Morgan Stanley.

Just one question about the autopilot software development. As you are putting the hardware and the software and learning capabilities in the entire suite of incremental production you're going to have lots and lots of very rich data that is going to be brought to you for analysis and processing and learning.

And the question is when, Elon, would you say would be the earliest reasonable opportunity for you, and perhaps backed by the scientific community and your own community in your Company, to make the case, to make a strong case to the regulators with the empirical data as you get it and analyze it of the safety of the vehicles, even if not in a fully autonomous application, but even in a semi-autonomous so that you can bring more visibility and transparency to the urgent need to address the spiraling death and injury on our roads? Thanks.

I should say we do work actually very closely on a daily basis, and have for a long time, with NHTSA and other regulatory agencies around the world really at a very detailed level. So they're certainly aware of the nitty-gritty, and as I've said before, we already see a significant improvement in safety with semi-autonomous features.

And what lessons about the outside is -- are early the cases where the version one of autopilot actually liked to mitigate the accident. So that the impact velocity went from being potentially fatal or severe injury to customer stepped out and walked away.

There are many of those which provides a much more statistically significant sample set than fatalities. Because fatalities are extremely rare, and you need really 1 billion miles or more to try to achieve a statistically significant conclusion on fatalities.

But as our fleet grows, and it's growing rapidly, the number of semi-autonomous miles grows to the point where I think we are now starting to approach 1.5 million, almost 1.5 million miles per day of autopilot through all kinds of road conditions and weather throughout the world. The more time that goes by the more miles we accumulate, the stronger the argument gets, but the confidence interval tighten and it becomes clearer and clearer. So I'm really quite optimistic about where things are and where they're headed on that front. I think they are headed to a good place.

Okay. Thanks.

Jeff Osborne, Cowen and Company.

Good afternoon. Just two questions on my end. One, how do we think about the cadence of CapEx? Should it in 2017, should it persist at a continued rate in the first half of 2017 up into the Model 3 launch at a similar run rate as you're seeing here Q4? Or what's the thought process there?

You will see a ramp up in Q2, in Q1 and Q2 as you would expect as we get closer to production. And then a lot of the payments come after we start productions in Q3, Q4. And there will be obviously expenditures on new vehicle development.

Expect it to ramp up a fair bit over time, but I stand by what I said earlier which is currently if we did not go out and raise a bunch of money, our current plan says we don't need to raise any money. It gets a little scary in terms of how much capital we have in the bank relative to our sales volume, but currently it's raised capital it's something it's nice to have, not a necessity.

And maybe it's a smart move to derisk things and all that. So just looking at the bigger picture, taking into account also that we're designing the 3 program to be a cash generator, that the faster the 3 grows the stronger our cash position. I don't think you need to worry too much about CapEx being a dilutive event or something like that.

Just so it's clear, what Elon is talking about is a step up from our Q3 levels, not a step up from Q4.

Do you care to throw a number out there for CapEx for 2017 at this point, or do you think of it flattish but front-end loaded on 2017 versus 2016? Too early to say?

Higher in 2017 than 2016 for sure.

Then, around CapEx as well, any thoughts on partner commitments to Gigafactory? What's the trend there? And then also I might have missed it, but what's the reservation count for Model 3? I missed that in the release if it was there.

We see very strong supplier commitments on Model 3. We don't see any deficit in supply commitments there. They're very strong. This is the most interesting vehicle program, maybe the most interesting product program in the world. And so suppliers really want to be a part of something like this.

As for the 3 deposit number, this is not something we comment on and not something that is a figure of merit in any way. We do no promotion of Model 3, we don't advertise, but we don't advertise in general, but we don't -- how often do you see me mentioning a Model 3? I think people sometimes forget that all we did for the Model 3 was a half hour webcast. There's no advertising, no guerrilla marketing campaign.

We sent out a few tweets like hey, there's going to be a webcast, and a lot of people decided they wanted to place a deposit for the car, which is cool. But we didn't want to get people too distracted from today's product in favor of tomorrow's product. And then when somebody comes into our store to buy a Model 3, we say well why don't you buy a Model S or an X instead?

So we anti-sell the 3. Still a lot of people ordered the 3, but whatever. Plus the 3, we're basically sold out the first year of production. So a good 12 months of production or thereabouts. So what's the point of trying to sell the 13th month of production? Very little to be had there in doing so.

Perfect. Thanks so much for all the details. I appreciate it.

David Tamberrino, Goldman Sachs.

Thank you. Just want to circle back on a couple of things said earlier. First on the autopilot, you mentioned that you worked very closely with NHTSA.

I'm wondering what your take is on the push from the recent document, the Federal Autonomous Vehicles Policy, that really is looking for data sharing among OEMs? I think you're probably clearly in the lead with vehicles on the road and miles per day of data that you are aggregating, wondering what your take is on potentially opening that up and sharing with some of your competitors?

We'd be happy to share information with our competitors that would help improve safety. Happy to do so.

Interesting. Then the second one is really just on the cost side, you think about a traditional OEM and their supplier relationships, there's typically annual price downs ranging in the 1% to 3% range, sometimes more for more commoditized products. And you're very vertically integrated. I wonder how you think about internal price downs and gaining economies of scale for the Model 3 and what you're really looking to achieve from an operational efficiency standpoint on an annual basis with the parts that you have going into your vehicles?

Model 3 efficiency as a whole, that really is a quantum change in productivity, like really crazy. I mentioned this before, but as we go to high volumes what really matters is the factory.

The machine that designs the machine. The machine that creates the machine is -- becomes actually of greater significance, much greater significance than the machine itself.

That's where we have most of our engineering team working on. So our internal code name for the factory machine that builds the machine is the Alien Dreadnought. The point at which our factory looks like an alien dreadnought, then we know it's probably right.

So we think with Model 3 will be Alien Dreadnought version 0.5 approximately. Then it will take us about another year or so, I don't know, summer 2018 to actually get to Alien Dreadnought version 1.

I'm a little bit hazy on quantifying crazy, is there any rule of thumb that you can point to with what you're looking to achieve, at least in terms of bringing the cost down from a component level from the S to the 3, not even thinking about the X given the increasing complexity that was involved with the vehicle?

It's a rough approximation. We think it could be about half.

Okay.

It's not like everything is half, some things are way less than half the cost and some things are more than half the cost, but on average about half.

Predominantly internally sourced?

Well, it depends on how you consider the value chain, but yes, I guess arguably it's the majority internally sourced. But there's still a huge number of suppliers. The thing that happens when you're -- once you start making almost all major subsystems internally, your supplier count actually grows dramatically. You have far more suppliers, not far fewer, but they're at the component level not at the major subsystem level.

Just one thing I'd add to that too, regardless of sourcing supplier, the way to think about cost, and this goes all the way back to first principles; what's the value of the commodities in the part? What does it cost to reasonably turn those commodities into a usable part? What's reasonable labor and overhead? And that's how we think about all material cost decisions, internal or external.

Our long-term aspiration for the machine that builds the machine, the factory dreadnought thing, is the long-term aspiration is limited physics. I'd maybe call it limited physics manufacturing.

We should be moving on to the next question.

Joseph Spak, RBC Capital Markets.

Thanks. I wanted to ask a question on the leasing. I know you pointed out that the percent of vehicles that are subject to the RVG this period I think declined by 4 points. And I don't know whether this was coincidental or not, but it looks like the direct lease percentage also went up by about 4 points. As you dwindle down the RVGs, are you planning that the ultimately straight is somewhere in that low to mid-30% range?

There's a bunch of different levers here. So one that is worth pointing out and we haven't talked a lot about is we've put out some very compelling loan products in the marketplace.

We're working with partners.

Different partners; through partners. So of course we always want to continue to do that, and we're always looking for ways where we can provide compelling and useful financing programs for our consumers. Whether they lease through a partner, whether that is a loan through a partner, or whether we leverage our own balance sheet in the case of a direct lease, we'll do that too.

Really it's about the consumer experience, and if we can use other folk's capital for that, great. If we use our capital for it, that is fine too and we're willing to make those decisions.

Okay. And then just back on autonomous, maybe to ask Adam's question a little bit different way, Elon, you talked about a cross-country trip in 2017, but in terms of turning it on for the consumer, I think in the past you said you need about 6 billion miles travel for regulatory approval. If I just do some crude math, based on your delivery time line that seems like at some point in 2018 you'll get there, maybe it's a year or so later if you believe consensus deliveries. But if you put aside the regulatory issues, is that roughly the timeframe you think it's ready for the consumer?

I think the timeframe that we think it is ready and then the timeframe that regulators will approve, you've got to present the data to them, they've got to think about it, and they've got to render a verdict and that can sometimes be a long process. And it can vary quite a bit by jurisdiction.

I think we may see some jurisdictions giving the okay a lot sooner than others. But when you think about the global average fatalities, it's somewhere around 60 -- one fatality every 60 million miles on a global basis.

So if you are at 6 billion miles you're 100 times the -- what the fatalities per mile. You really start to get quite statistically significant at that point. And it can make quite a strong argument I believe at that point that it would be morally wrong not to allow autonomous driving.

Thank you.

We're coming up on the hour mark. We have one other analyst on the call, and then we have some journalists we definitely want to hear from as well. So let's -- you want to go a little bit over an hour here, Elon?

Sure.

Chanel, let's have the next question, please.

Charlie Anderson, Dougherty and Company.

Thank you. I'll just ask one question. There was a reference to the Tesla network and the ability to buy self-driving today. So I wonder, Elon, if you could talk maybe philosophically about how you are viewing Tesla network? Is it something that will generate income for Tesla to help develop future products, et cetera, at a reasonable gross margin? Or is it something that you'll use more for market share gain to help people offset the price of the car long term? Thanks.

Okay. I think it's a bit of both, really. This would be something that would be a significant offset on the cost of ownership of a car, and then a revenue generator for Tesla as well. Obviously the majority of the economics would go to the owner of the car.

Sometimes it's been characterized as Tesla versus or Uber or Lyft or something like that. It's not Tesla versus Uber, it's the people versus Uber.

All right, Charlie?

Thanks so much.

Daniel Sparks, The Motley Fool.

Thanks for including us in the call. I just wanted to get a little perspective on, I noticed in the shareholder letter, the narrative kind of shifted. In Q2 you were saying that you're aiming toward volume production toward the end of 2017.

But now the letter is saying you're looking for volume deliveries in the second half of 2017. Am I just reading into this too much, or does that reflect a greater confidence on Management's part?

I think our confidence has been practically the same. Obviously as time goes by there is some amount of the uncertainty is collapsed, and so I guess you could call that confidence. But yes, it's looking good for production volume second half of 2017.

As always, I really want to remind people that a car consists of several thousand unique items. We can only go as fast as the slowest item. And so what we're trying to do in advance of 3 production is to increase the scope of Tesla's internal capabilities so that we are internally capable of making almost anything.

Kind of like reserve troops. You don't know exactly where they will be needed, but it's a good idea to have them. And so that we can minimize the degree to which a single supplier can stop the entire production line.

Okay. Great. And then as Model S and Model X with higher levels of sales recently, high levels of deliveries, and as Model 3 approaches, do you feel confident in these levels as Model 3 approaches? I know that we haven't talked too much about 2017, but just speaking as far as trajectory for those deliveries go and how we could think about it?

Another thing I really want to emphasize is the production ramp tends to look like -- is exponential. Ultimately it is an S-curve. An exponential goes to linear then goes to log. And it's very difficult to predict exactly where that beginning part of the exponential and the S-curve fits in between quarterly reporting.

A shift of even a few weeks one way or the other can have quite a dramatic effect on what it looks like in that quarter, but that's not indicative of the future. So we're telling you what we -- we're giving you the best assessment we have short of having a crystal ball.

I think things will look very good exiting 2017, but it will be complicated and bumpy and dealing with a lot of unexpected issues in the first -- at the beginning of Model 3 production in Q3, Q4. Q3 particularly is very uncertain because it is the beginning of an exponential. It gets pretty clear in Q4, then it starts to be really crisp in the Q1/Q2 timeframe of 2018.

All right, great. Thank you.

Tim Higgins, WSJ.

Thanks for making time, I appreciate it. Just to go back to the capital issue, I hear you saying you don't need to raise capital this year, and I hear that you probably won't do it in the first quarter of next year, but what about next year in general? Should we look at that as a second half or first half event? You want to raise capital in the first half of next year, even if you don't need it?

I think we cannot make -- actually I don't think it's legal for us to make specific predictions of certainty in respect to doing an equity raise or something like that. It's really exactly what I said before, which is our current projections, and this should probably be taken with a grain of salt, current predictions say we don't need to go out and raise a bunch of equity. There could be unexpected negative things that occur, there could be some global macros, economic slowdown that could be -- who knows what could happen. And so there may be value in derisking the business and just having higher capital reserves. We're not ready to make that decision yet.

Okay great. Thank you.

Phil LeBeau, CNBC.

Hello, Elon. Quick question, in your shareholder letter you mentioned that you are continuing to explore possibilities for expanding production to Asia and Europe. As you start to look at the production ramp and expanding your facilities in Fremont, do you have a timeframe for when you might make a decision in terms of, I think this is when we will probably make some kind of a decision about another production facility, whether it's in China, whether it's in Europe, wherever it might be somewhere beyond Fremont?

Right now we're really focused on Gigafactory 1 and Model 3. We're spending very little time on facilities outside of Fremont, California and Sparks, Nevada. So it's really hard to say at this point except to say it's pretty obvious that long term you want to have your production close to your consumption so don't have massive logistics costs transporting cars halfway around the world.

And yes, so that's -- I think we're probably not ready to talk about that now and we just don't have a fully formed idea now. Probably end up talking about that next year.

Great. Thank you.

I'm showing no further questions on the phone lines at this time.

All right, thanks a lot, Chanel, and thank you everyone for joining us today.

Thanks, everyone.

Ladies and gentlemen, thank you for participating in today's conference. This concludes today's program, you may all disconnect. Everyone have a great day.