ARK给出的2031年特斯拉基准情形估值为4,400美元，约为当前股价的12倍，这是他会满意的结果；其中robotaxi占2031年估值的81%，新纳入的Optimus占14%。最大风险是监管延迟导致robotaxi放量放缓，不过ARK表示这类延迟正在减少。

$4,400 per share. That's Arc Invest's expected value for Tesla stock in 2031 in the base case. This would be more than $17 trillion in market cap, about 12 times where Tesla is today.

As I record this, just under $369 per share. The new model dropped just a few hours ago. There's an accompanying blog post. We're going to be looking at some of the key highlights in particular. what's changed between the last model and the new model and the most important biggest drivers of value.

This is a 5year model horizon and ARK's base case is well above a 10 times increase in Tesla's stock over that time frame.

Now, the headline number $4,400 per share. There's also a bare case. Surprisingly, it's not that far off where the 4,400 base is at just 3,300.

I want to keep this video pretty concise, but for those who don't know, the way ARC models this is a ton of different simulations of potential outcomes. They then essentially line those up from worst to best.

Right in the middle of those is where the base case is, the 4,400 per share. If you divide that span of outcomes evenly, this is at the 50th percentile. Right down the middle, go halfway toward the worst case.

We're at the bare case. The 25th and 75th percentiles are the bear and bull case.

The bear is 3,300 per share. Still roughly a 9 times increase from where Tesla is today. The bull case not that much above the base. This to me suggests that the range of possible outcomes has tightened.

Meaning ARK's confidence in their estimates, their analysis, their model has increased.

The full ARC model has 32 inputs that meaningfully affect the valuation.

First of all, ARK's most recent model released in 2024 was a 5-year model. So, it went out till 2029. And the estimated fair value for Tesla then, or their expected fair value was $2,600 per share.

So, this is a whopping, are you ready for it? 69% nice increase by looking 2 years further into the future. The main driver, no pun intended, of this huge increase in valuation by just extending a further two years is the impact of robo taxi.

This is to be expected. Scaling robo taxi doesn't happen instantly. So the further out into the future you look until Tesla's robo taxi business is at massive scale, the more substantial the revenue and profits from their ever larger robo taxi business will be, meaning the higher the fair valuation for the company.

So expect this trend to continue. the next model when it extends a year or so into the future, another big bump from even more robo taxi revenue and profits.

And don't get me started on Optimus. So, the majority of the increase from 2600 to 4,400 isn't necessarily major changes to the underlying assumptions. It's just literally, well, the model's now looking at 2 years of additional robo taxi scaling and the impact on Tesla's business.

We also have an expected value. This is slightly different from the base case, which is literally just the 50th percentile. Sorry for the people who aren't super nerdy. Bear with me.

We'll get to the juicy [ __ ] soon. This is the last, I promise. For those of you with at least a touch of the tism, though, this is their weighted probability weighted 2031 valuation.

Not too far off the 50th percentile mark, but this is a slightly different methodology. Very similar end result.

Also true of the bear and bull case. They're not too far off, you know, within about 10% of their respective 25th and 75th percentiles. Part two, the deep dive. Oh, by the way, in case you guys haven't noticed, got a little bit of help from my favorite product of all time, Grockbot, on this presentation.

So thanks Crockbot. Couldn't have done it without you. We're now looking at a deep dive. These are the four things that either move the valuation most or separately are just really worth taking a closer look at.

So Robo Taxi accounts for 81% of Tesla's 2031 enterprise value. Absolutely carrying the company.

What's really worth looking at is that the addition of Optimus, which wasn't in Ark's previous model, is now accounting for 14% of Tesla's 2031 enterprise value.

And ARC is still quite conservative on their Optimus assumptions. Again, this is a 2031 5-year time frame. So, it's not that far into the future. Optimus is going to take a long time to hit insane scale.

Even so, Optimus already at a 14% share of the enterprise value in 2031.

So this isn't as much a factor of robo taxi itself being dramatically reduced in terms of its impact financially on the business. It's just the addition of Optimus as a much larger chunk to the overall pie.

And Optimus again the assumptions aring is still quite conservative. Yeah, imagine that arc conservative. The point I'm getting at, when ARC drops their new Tesla valuation model 5 years from now, looking another 5 years into the future, there's an extremely high chance that Optimus is going to be looking like what Roboaxi was previously in terms of dominating the valuation.

Not cuz Robo Taxis has done, but Optimus, the scale of that opportunity is just going to be so much greater than Roboaxi could ever be.

In 2031, the Bearbull and base case for total Tesla robo taxis, the bare case is at 4.8 million. The base case roughly doubled that at 8.1 and the bullc case barely above the base to be honest around 10 million.

The net revenue from right hail then in these cases bear a mere half a trillion or so. Base case about 3/4 of a trillion and bullcase not far off $1 trillion.

Of note, Tesla hit 1 million driverless miles in around 7 months. Whimo pulled off the same feat in 27 months.

Tasha Keiny upon this announcement wrote on X, quote, "Robbo taxi still comprises the majority of value over the next 5 years, though Optimus sales inflect more here than previously modeled.

Energy storage in ARKs model by the way, a whopping roughly 2% of 2031 enterprise value, which I think in finance jargon is [ __ ] all. By the way, they're still talking about 100 plus billion dollars of revenue, but relative to the other businesses, not a needle mover. so insignificant.

You could literally delete this and the valuation would barely budge

onto Optimus. Arc estimate 1 million Optimus robots would be sold to outside customers by 2031. And just to clarify, these are Tesla Optimus bots. In the past, Elon has made some estimates about the total number of humanoid robots that there'll be at certain points in the future.

Not just Tesla robots, but total robots. Tesla obviously want to aim for a very large share of that. But to be clear, ARK expects 2031 1 million Optimus robots sold to customers and presumably many more built and retained for Tesla's own needs.

Their base case for Optimus revenue $212 billion, which even at that stage would be in excess of their bull case for Tesla Energy.

An overview of Optimus in ARKs model. Depending on the scenario, outside sales start between 2027 and 2030. They expect the Optimus robots will achieve human level skill around 2031, also requiring 4 GW of compute.

The Optimus Academy, which Tesla previously disclosed, it's basically play school for robots to learn how to do things, manipulate objects, could have 20 to 30,000 robots trained inhouse at Tesla.

The Ark model also actually includes a line item for Optimus capability increase. Interesting way of thinking about this. ARK expects that in 2031, just that year, Tesla spends around $225 billion on AI infrastructure and chip development, and that 10 to 25% of Tesla's year-end cash goes to AI infrastructure.

They also expect that the initial research arm of Terapab, which for those who don't know, the plan is build a little baby Terraab to test stuff out, to innovate, to invent new ways of chip fabrication at relatively small scale, see if it works.

If it does, that knowledge learning IP technology then gets scaled up. And when I say scaled up, I really mean scaled up to the near citiesized monstrosity that will be the full final terapab, the research terraab, maybe $3 billion of cost.

SpaceX likely to foot the bill for the majority.

This is a cool little slide. Grock put this together obviously after analyzing the model. This is what actually impacts the share price the most. Think of these like a bunch of different levers.

If you pull one of these levers, how much of an impact does it have?

And at this point, robo taxi far and away the biggest driver of value just behind it is the actual fleet size. And obviously this makes sense. The more robo taxes you have operating, the more money you can print.

Next is a bit of margin. Essentially, how profitable is the business itself? That actually meaningfully affects the valuation. Investors typically tend to pay more for a highly profitable high margin business.

AI spending also has a big impact. And here's why. Even though this is a cost, ARK's assumptions are that the more money you spent on AI, the more value is created, hence the more valuable the company.

Cars sold a moderate impact. And then we get into the fairly small numbers. Optimus, energy storage, not massive needle movers, at least not in this horizon.

So here's what's been added versus removed. And quick caveat, by the way, I'm assuming that Crockpot got its homework correct. Haven't checked this yet. That's how much trust I now have in my Crockpot, by the way.

So this is a comparison between the previous model and the new model. What's been added? Optimus robots being sold to outsiders customers. In other words, AI and terraab spending, willingness to pay pricing, which is a whole super nerdy semi- autistic part of the model itself regarding robo taxis plus Elon stock-based compensation, dilution, and the total share count.

The finance nerds will understand everybody else. No need to worry about it.

What's been removed was the human-driven ride hail business. I believe previously ARK was modeling that they were literally going to be human drivers much like Uber operating at some scale before Tesla massively scaled robo taxis. Not anymore.

Tesla insurance in the in the bin. And actually a side note on Tesla insurance. I have learned as a midwit content creator that scaling insurance products globally is unfucking believably slow, arduous and painful.

Desert insurance is a great product and it scaled actually out quite quickly in certain areas. I think China was actually the first market ever years ago, but the roll out has been much slower than I would have realized.

And this is obviously due to a bunch of regulatory burdens, hurdles, red tape, bureaucracy, and a bunch of other [ __ ]

Marks also removed China Robo Taxi from the model entirely. And now that Robo Taxis have actually launched, obviously they don't need the launch year input for Robo Taxi.

Now, these are the biggest changes versus the previous 2024 model. We saw the expected value per share. This is the weighted version. 2,617 up to 4,342.

Robo taxi share of the total value was 88%. It's now 81%. Again, not because robo taxis are not going to meaningfully impact the business. It's just Optimus is now a massive slice around 14%.

Now, don't get me wrong, some of the robo taxi assumptions have changed as well, but the real big factor here, most of the difference is Optimus being added, not robo taxi being downgraded.

the miles per road best taxi per year 20 24 model 70,000 to 120,000 which I'll be honest obscenely optimistic even 70,000's pushing it now don't get me wrong it's possible a typical taxi driver full-time might do around 50,000 maybe you know in a really good year closer to 60 70,000 it's been adjusted now the new model to 65,000 so it's on the higher end but it's definitely tangible viable and credible remember an autonomous vehicle can be operating more than full-time hours this works out to be 178 8 m a day, 365 days.

If you're doing 12 hours a day, less than 15 m an hour. So, actually quite quite doable.

They've brought down their vehicle gross automotive margin cap. Previously, the range of simulations had gross automotive margins as low as 20% and as high as 80%. On the low end now, it goes down to 12.5, but on the high end, it's capped at 50%.

And they've also got the vehicle and energy storage multiple from 13 times to 10 times. So, a lot of these have come down.

and energy storage capex per megawatt. How much it's basically costing Tesla come down from 10 to 45,000 to 5 to 10,000.

And curiously, the Tesla valuation model just so happens to dedicate some space to SpaceX. Check this slide out. An actual screenshot. SpaceX plus Tesla revenue and EBIT, aka profit by business line 2031.

Just to orient you folks around this, this isn't Tesla on one side and SpaceX on the other side. This is Tesla plus SpaceX. This is total revenue by segment. And this is net EBIT.

The SpaceX AI business, the big pink slice, roughly half of the entire pie, 3.8 trillion total revenue. This is 1.9 trillion of that. Just abs fuckingutely dominating. Starink, a big chunk here as well, 450 billion.

Tesla's robo taxi business, around 3/4 of a trillion in the dark navy. Optimus in blue, energy storage, the darker green. and the electric vehicle business in purple.

Of this entire $3.8 trillion total revenue pie, SpaceX $2.35 trillion and Tesla 1.45. SpaceX accounting for more than 60% of the total revenue.

However, when we look over at EBIT, essentially profits 2031, SpaceX AI is absolutely dominating. So much so that in this chart, AK has combined Tesla's entire [ __ ] business into one segment.

And it's still dwarfed by just the space XAI AI business alone in pink. Starlink is green.

So this is the stuff that really matters. Revenue irrelevant. Profits that's what really matters. And what do you know? The SpaceX share of net 1.9 trillion represents 69%.

To reiterate, ask estimates for how profitable these companies will be, how much money they're actually printing, how much profit they're making in 2031. a total between the two companies of $2.74 trillion.

But of that, SpaceX represents 69% $1.9 trillion. So 69% one nice two that's a big chunk not far off 3/4 of the profits of the combined EBIT of Tesla and SpaceX according to Arch Invest in 2031 accounted for by the SpaceX business and of that just SpaceX AI is the vast majority.

Now, this is interesting to see an ARX model for two reasons. One, they obviously expect at some point Tesla and SpaceX merge. And two, they've obviously started thinking about, well, if that is to occur, what would be the fair ratio in the event of a merger?

The way you figure that out is what company, if you think about the new combined entity being worth 100% of the total value of that, who's bringing the most to the table? Who's contributing the most value?

What's the fair ratio? This is what all Tesla and or SpaceX investors should be doing. themselves. You got to take the time to try and model out both companies over the long term.

I go much further than 5 years, but the merger horizon potentially probably is within 5 years. You got to try and figure this stuff out.

What it's worth, my valuation model, I've actually just asked Rockill to pull the numbers. I have an even higher share of SpaceX in terms of EBIT in 2031 about 10 or so% higher than ARC's estimate of 69%.

And the reason you need to think about this is because if there is a merger proposed by these two companies, you got to know whether or not what's being proposed is a reasonable and fair merger ratio or not, you don't take the time to do this exercise and figure out well, you know, what percent of the potential future profits and value would be the SpaceX AI business, Starink, and so on, you're flying blind.

Interestingly enough, I launched my 20-year Tesar and SpaceX valuation model on Patreon a couple of months ago. If you're interested, check out the link below. It's available to supporters at investor level and above.

But even those who are entry- level supporters get a look at the snapshot which does include the funnily enough the merger ratio kind of end destination of this modeling of which companies worth what at some point in the future we're about today.

Not to get too sidetrack but a fair ratio of a merger. It's not just a matter of which company's more profitable at the time of a merger but rates of growth and blah blah blah.

So there's a lot more to it. In fact, my model, which I think has about 50 [ __ ] sheets in it, also includes a fair ownership percentage, which is really the bottom line ratio in the event of a merger.

And I've just pulled this up now. Ask Grock Build in the latest version of the model. The fair ownership ratios. And this is fascinating. My model has the fair ratio at a merger.

SpaceX accounting for 66.5% of the value of the combined entity, exactly almost exactly double that of Tesla.

And recall we just looked at ARC's estimates for 2031 EBIT where SpaceX was accounting for around 69% actually 69% of the profits.

Again, you want to see the full valuation model, just check it out on Patreon, available to people at the investor supporter level and above. Or go look at ARX. That one's free.

It's actually quite interesting though, just to elaborate on this a bit more. You can't accuse me of copying Arcs homework cuz my model was published months before ARX just dropped there.

But it is interesting that at a fairly high level, Ark and I seem to be somewhat in the same ballpark in terms of which company in the future, SpaceX versus Tesla, could be contributing the most value and or worth the most.

I guarantee you most Tesla investors who haven't actually taken the time to model out SpaceX would think that's [ __ ] What the No.

On to the risks. The biggest, of course, slower robo taxi ramp due to regulatory delays. The next would be Elon's sudden departure. I think this is an extremely unlikely outcome. The next natural disaster or a scam demic.

Though, Ark notes that regulatory risks quote seem to be diminishing. It's still a robo taxi story. Robo taxis accounting for 81% of the value.

Optimus now in the math. And while it's not completely shifting the valuation, we've seen that even with just 1 million Optimus robots. And remember, Elon is on the record saying that he expects longer term there'll be 100 million plus humanoid robots, maybe a billion plus, maybe 10 billion plus, not all Teslas, but Tesla's aiming to take a huge slice of this market.

And even with just 1 million Optimus robots, it's starting to move the needle in terms of Tesla's fair valuation.

And number three, expect Tesla to spend approximately 12 gigaf [ __ ] tons on AI.

Quick overview of ARX $4,400 per share expected value for Tesla in 2031. I'd be pretty happy with that, just quietly.