The stock of Tesla (TSLA 0.49%) was up almost 700% in 2020. As the largest holding in Cathie Wood's ARK Innovation ETF (ARKK 1.84%), its massive price spike helped Wood's fund gain 152.8% in 2020, catapulting Wood to investing superstardom.
Even down some 25% in 2021, Tesla's stock still trades at a whopping $640 billion market cap and 23 times sales, which is very expensive for an automaker. But that's not stopping Ark Invest from putting out a brand new $3,000 price target by 2025, up from $660 today, with a $1,500 "bear case" and a $4,000 "bull case" -- and its overall target skewed more toward that bullish scenario.
That certainly raised some eyebrows, including mine. So I dug into some key assumptions behind that bull case, which Wood published on Github and Ark Invest's website. I found many of the assumptions behind that figure to be a stretch... and that's putting that mildly.
Selling as many cars as Volkswagen, but at higher prices and higher margins -- even after price cuts
First, Wood's "bull case" model forecasts 10 million Tesla vehicles sold in that year, up from about 500,000 delivered in 2020. It's going to take a lot of capital to get there, and Tesla has big incoming competition from the likes of Volkswagen (VWAGY 0.38%), which just announced an ambitious plan to ramp up its EV capability over the next few years.
Speaking of Volkswagen, it sold the most vehicles of any automaker in 2019 at nearly 11 million units. So, Wood's assumptions for Tesla are no doubt aggressive, but it's possible Tesla could get there over time -- though by 2025 is a stretch.
There are quite a few complications on that road to 10 million. Wood outlines Tesla selling 10 million vehicles at an average price of $36,000 per vehicle, which seems aggressive. Volkswagen's average sales price per vehicle in 2019 was a little more than $27,550 per car. Obviously, price matters when talking about mass-market car volumes, which need to be affordable to move that many units. Could Tesla find 10 million buyers at that higher price point? It's still unclear.
Meanwhile, Tesla will have to bring down costs massively over the next five years to make that happen. Last quarter, Tesla's average price per delivery was about $51,553 per vehicle. So, we're talking about Tesla lowering its average selling price by 30% over the next few years and still being more profitable than every other mass-market carmaker out there.
Yes, Tesla will be delivering cheaper Model 3s and Model Ys by then, which have much lower price points than the original Model S and X luxury models. However, even the Model 3 currently starts at $37,500, and that's just for the standard, bare-bones version. Therefore, to think Tesla will be able to lower its prices to a company average of $36,000, all while generating 25% vehicle gross margin (up from 24.1% last quarter) also seems like a huge stretch.
So, Wood thinks Tesla can sell as many cars as Volkswagen, even though it will be a more expensive car. And even though it's more expensive, it will still take an average 30% drop in average selling price to even get it to that level. All while margins increase simultaneously.
It's possible Tesla could accomplish all of that, but it's a long shot to say the least. But the vehicle sales forecast is probably the least fantastic assumption Wood makes.
Selling auto insurance at 40% EBIT margins
Another pillar of Wood's case, even if it's a small part of the value, is Tesla's getting into auto insurance in a bigger way. Wood models Tesla's ability to better insure its own vehicles because of all the massive driving data it has on its customers. Wood forecasts that Tesla could insure 40% of its own vehicles by 2025, earning $23 billion in revenue and earning a whopping 40% operating margin on that revenue. Ironically, that's for the bear case. In the bull case, Wood expects the insurance revenue to roll into an autonomous robotaxi fee (which I'll get to in a minute).
A 40% operating margin in insurance is seems pretty wild. Progressive, a top large-scale auto insurer today, made a 13% operating margin in 2019. And Tesla isn't exactly the only company using lots of driving data to better underwrite insurance. A whole slew of new insurance upstarts are using big data to underwrite car insurance, and you can be sure the legacy players such as Warren Buffett's GEICO, with huge resources and smart people, are also using lots of driving data to underwrite car insurance as well as anyone.
In other words, I don't see where Tesla's data advantage on its own vehicles would allow it to do better than companies that specialize only in insurance. Today, Tesla only offers insurance in California, and even in that one state, it outsources underwriting to insurance partners. The fact that Wood thinks it can get to $23 billion insurance revenue, begin underwriting all the policies itself, and also become more than three times as profitable as the best car insurers today also seems completely unrealistic.
Launching a human-driven robotaxi service (before going autonomous)
Perhaps the weirdest assumption -- actually, a recommendation by Ark's analyst to Tesla -- is that Tesla start its own human-driven ride-hailing service now, competing with established leaders Uber Technologies and Lyft. The thinking is that a present ride-hailing service would lay the technical platform and groundwork upon which to launch Tesla's autonomous robotaxi service down the road. As we'll see soon, that autonomous service accounts for over half the company's market value in the bull case.
But first things first. How is Tesla supposed to launch a human-driven taxi service? The whole point of Uber and Lyft is that they recruit drivers with their own cars, no matter the brand. You can't just recruit only Tesla owners for a ride-hailing service. If nothing else, I'm guessing the average Tesla driver already has a job and isn't going to want to drive people around.
So, is Tesla going to give drivers expensive Teslas to drive customers around in? If so, that would eat into vehicle sales. If they sell Teslas to these drivers, the company would likely have to discount them pretty heavily. And they'd still be a late-mover against Uber and Lyft in an industry that benefits first-movers.
The human-driven ride-hailing service actually isn't in the 2025 bull case, as it's meant as a stepping stone to the fully autonomous service. However, it is a significant part of the bear case at $1,500. Wood estimates Tesla sells 40% of its production into the service (at full price), and that the service generates $42 billion in revenue at extremely high margins. For perspective, Uber only made $14 billion in revenue in 2019, at about 14% adjusted EBITDA margins. So to think Tesla will become three times as large as Uber with much higher margins, all while not eating into its own vehicle sales, and all within five years, seems fanciful.
And the whopper: An autonomous robotaxi service by 2025?
The biggest factor in that enormous $4,000 bull case scenario is the likelihood of a fully autonomous global ride-hailing service in 2025. In that year, Wood forecasts a stunning $327 billion in autonomous ride-hailing revenue, at an extremely high margin -- probably around 75%. And that's in addition to $367 billion in full-priced vehicle sales.
Ark's thinking is that Tesla currently has an advantage in autonomous technology, because of the visual data it gathers from its current fleet and leading artificial intelligence capabilities. It's not totally out of the question that Tesla could surpass Waymo, as well as all the other companies racing to get an autonomous vehicle service out on the road, but it's far, far from certain. Ark's 50% probability that Tesla does all this by year-end 2024 is implausible at the very best.
Not only are there technology hurdles to clear, but also legal ones. Given the state of politics today, I find it almost impossible that a global, large-scale autonomous ride-hailing service would be approved by governments by 2024. Several years ago, many of the autonomous leaders today -- Tesla and Waymo included -- had previously forecast we'd already have these autonomous vehicle services by now. However, beyond a few pilot projects in small geographies, they're nowhere to be found.
But there's another glaring hole in the thought process here. In ARK's model, Tesla's autonomous ride-hailing service revenue is all supposed to be incremental to vehicle sales. In other words, not only will Tesla sell all these cars, but it will get the autonomous revenue, too.
One wonders how that would work. Would people buy their own Teslas, then rent them out when they aren't driving them? I don't think many Tesla owners are going to want strangers using their cars all the time. And even so, wouldn't there be diminished supply during the day, when most people use their cars, then an oversupply of vehicles at night?
Even if this super-rosy scenario in which Tesla launches successful autonomous technology and government regulation happens in a timely manner, I'd assume people would either purchase their own Tesla or use the ride-hailing service, not both. As such, I think there's an awful lot of double-counting going on in that bull case model.
Check yourself before you wreck your Tesla model
Tesla is still a great company that has brought electric vehicles into the mainstream, and has achieved really impressive things. It could even one day grow into its current market cap, if a lot of things go right. However, even the $1,500 price target (based on a human ride-hailing service) seems dubious to me. And the $4,000 target has not one, but several out-of-this-world assumptions contained within it.
Tesla's stock has been a huge winner -- admittedly, much more of a winner than I thought possible. So, you can take my skepticism with a grain of salt. However, if you really believe Tesla will reach $3,000 in a few short years, make sure you understand and believe the assumptions behind that valuation, and aren't just buying into the scenario because one newly famous investor said so.