The Irreversible Decision That Could Bankrupt Tesla
Look, a self-driving squirrel!
On April 5, Reuters ran a scoop that was bad news for Tesla investors.
"Exclusive: Tesla scraps low-cost car plans amid fierce Chinese EV competition," the headline said.
Citing three sources and internal company messages, Reuters reported that plans for Tesla's long-awaited Model 2 — the mass-market car that was supposed to be built in Texas, launch in 2025, and cost just $25,000 — had been canceled.
The new plan, per Reuters, was to use the Texas plant for self-driving robotaxis instead.
Elon Musk, who never lies or makes stuff up, then accused Reuters of lying... and a few hours later seemed to validate the scoop.
Musk's initial response to the Reuters story was to say, "Reuters is lying (again)."
But then, a few hours later, Musk tweeted that Tesla's robotaxi would be unveiled in August.
Haha, oh boy. We have painfully obvious questions.
First, why would Reuters — a global news operation that was founded in 1851 — decide to risk more than 170 years of credibility on a "lie" that could be disproven within minutes of the story being released?
All Tesla would have to do is say "we are still on track for a 2025 Model 2 launch target," maybe give some color with factory floor pics or designer plans, and that would be it.
Second, if the Reuters story was false, why would Musk say "Tesla Robotaxi unveil on 8/8" literally within hours of the first tweet accusing Reuters of lying?
I mean, who knows, right?
But it sure looks like the Reuters scoop was correct based on the pairing of Musks's responses.
If the Tesla Robotaxi is getting an unveiling in August, and the Model 2 is not, and 2025 is only eight months away... that looks like a company pivot away from the Model 2 and back toward the robotaxi, which is exactly what Reuters reported.
It also looks like Musk still has zero control of his tweeting, er, Xing impulses after all this time.
Accusing a news organization of lying when they probably have the receipts, and the actions of your own company probably prove them right, is, uh, not smart.
Our hunch is that Musk's hours-later post regarding the robotaxi unveiling was a way to try and undo the damage of the first accusation, while simultaneously calling attention away from the scrapped Model 2 (no $25,000 Tesla on the horizon).
In other words, keep your eye on the self-driving squirrel — a robotaxi — not the painful revelation that Tesla's plans for creating a low-priced EV were so far from reality they were tossed into the bin.
But hey, cool, this means the robotaxi is coming! Awesome, right? Haha, sure.
To quickly revisit a history of Tesla "reveal" dates relative to launch times:
The Tesla Semi, an electric Class 8 semi truck, was revealed in November 2017 with a promise of deliveries in 2019. The actual first delivery was December 2022, just over five years from the release date. As of year-end 2023 around 100 Tesla Semis were built, with mass production reportedly still in the works.
The Tesla Cybertruck was unveiled in November 2019 (just before the Tesla Semi was supposed to launch, but didn't). Cybertruck deliveries began in November 2023, four years later. Tesla did not reveal how many Cybertrucks were built and delivered in 2023; estimates suggest fewer than 1,000 were delivered in first quarter 2024.
The Tesla Roadster, a sort of Tesla supercar concept, began taking pre-orders in 2017 with a hefty $50,000 deposit. The launch date of the Roadster was then pushed back again and again... with the latest estimate being 2025. (For those who put down $50K in 2017 hoping 2025 will be the actual year, good luck with that.)
So in terms of the window between "reveal" date and actual launch, we are looking at five years.... four years... and an estimated eight years and counting (the Roadster), with no meaningful production volume in the first two instances.
Using this track record rule of thumb, one might expect a 2024 robotaxi reveal to correspond to an actual robotaxi launch date somewhere around 2029.
Except, in our view, the more realistic robotaxi launch date will be "never."
The Tesla Roadster, while still a phantom product, could at least be produced in theory. The robotaxi core concept is based on technology that doesn't exist — and for Tesla may never exist (in terms of doing what it says on the label).
This takes us to the one decision that could potentially bankrupt Tesla. Or rather, there are a lot of things that could bankrupt Tesla, but this single bad call looms above them all as an existential risk.
Lidar stands for "Light Detection and Ranging." In self-driving terms, Lidar is like an extra sensory array for the car.
In the manner a bat sends out echolocation signals to navigate via sonar, Lidar arrays use lasers that bounce off nearby objects, enabling a vehicle to map its surroundings more accurately.
Most self-driving car companies — Waymo, Cruise, and so on — use Lidar arrays. Tesla does not.
Tesla cars rely on visual cameras only for self-driving capability, and it may be too late to turn back.
Simply put, if a lack of Lidar capacity turns out to be an Achilles heel in the self-driving race, that could be the end for Tesla.
Elon Musk himself is on record saying that "solving full self-driving" is life-or-death for Tesla, calling it "the difference between Tesla being worth a lot of money and being worth basically zero."
But here is the thing: When Musk rejected Lidar for Tesla vehicles years ago, he didn't do so because visual cameras are an adequate self-driving solution. He did so because he had to.
By that meaning, in the 2010s Lidar technology was simply too expensive to include in the cost of an electric vehicle (EV) being sold to the public.
In 2015 a popular Lidar array from Velodyne, a Lidar manufacturer, went for $75,000. You can't sell an EV that is already expensive with another $75K tacked onto the price for Lidar components.
Expensive Lidar was not an impediment for other self-driving car companies because selling cars was never part of the business model. Google's Waymo could use a $75,000 Lidar array, for example, because the expectation was not to sell cars to the public, but rather to develop car fleets.
Then, too, early users of Lidar knew the cost would fall dramatically over time as Lidar production scaled up to higher volume. In technology this is known as the waterfall effect: As an expensive technology gets cheaper, the unit volume explodes until, finally, the same technology is low-cost and used everywhere.
Musk couldn't wait, though. Tesla had to sell cars in the 2010s, and it also had to have skin in the self-driving game.
So Musk decided — out of necessity, as a function of the business model of selling cars — to say that Lidar was unnecessary and visual cameras would be fine. It was a reverse-engineered justification for the fact that Lidar was too expensive.
"Lidar is a fool's errand," Elon Musk said in 2019. "Anyone relying on lidar is doomed," he added, saying they were "expensive sensors that are unnecessary."
That argument was nonsense from the start. Everyone knows that expensive technologies tend to grow cheaper over time (via the waterfall effect), and self-driving cars need all the sensor help they can get to reduce the risk of accidents.
Years later Musk was banging the same drum. In 2022 he wrote: "Humans drive through the eyes that see, and a brain made up of biological neural networks that analyze the information. There's no reason an autonomous vehicle would not work the same way, with silicone neural cameras and networks processing the information."
That argument is ridiculous. First of all, humans are allowed to have accidents in a cultural sense; self-driving cars are not. Second of all, the human brain remains the single most powerful piece of technology in the known universe. If artificial intelligence has any hope of replicating what the brain can do, it may take a nuclear power plant worth of energy to come close.
And yet a self-driving car is supposed to be fine with visual sensors only because any day now the processing power of a human brain will be laying around?
Then, too, if self-driving cars are ever going to truly eclipse human drivers in terms of handling edge cases — like driving through a snowstorm — they will almost certainly require Lidar to do it.
Lidar arrays are no longer expensive. The cost is reasonable and they are getting cheaper by the day (the waterfall effect at work).
And yet, it is likely too late for Tesla to adopt Lidar in part because Musk has railed against Lidar for years, but more importantly because all of Tesla's production lines, all of Tesla's self-driving software, and all of the existing Teslas on the road are aligned to camera-only visual arrays.
This means that, if Tesla did try and switch to Lidar, it would have to retool production, reconfigure the software, and admit that all of the Teslas on the road that were still camera-only were inferior, causing resale values to plummet through the floor.
Keep in mind too that, as of this writing, Mercedes has SAE Level 3 autonomy certification in the United States and Tesla does not. Level 3 is a level of self-driving certification that no other automaker has, including Tesla. And does Mercedes use Lidar? Of course.
The means of throwing Tesla into a crisis tailspin would be a government regulation requirement that higher levels of self-driving certification require Lidar, non-optional, as part of the technology package.
This would be comparable to aviation regulations that say a certain type of backup safety system has to be installed on a commercial airplane before it can accept passengers.
As such, if regulators in large enough markets (e.g. California, Germany, et cetera) ultimately decide that Lidar is a must-have, not a nice-to-have, with cameras-only insufficient, then Tesla is not just in trouble... it is arguably dead. A "zero" in Muskian terms.
Is this outcome likely (regulators eventually mandating Lidar for all self-driving systems)? We would say yes, with odds higher than 50%.
After all, why wouldn't advanced sensor technology be better than a lack of such, especially when it comes to edge cases (snowstorms, et cetera) and safety?
And so, as the result of a fateful business decision in the 2010s — rejecting Lidar as incompatible with public vehicle sales — Tesla could be due for more than just a massive share price haircut. It could be doomed.
Until next time,
Justice Clark Litle Chief Research Officer, TradeSmith
TradeSmith is not registered as an investment adviser and operates under the publishers' exemption of the Investment Advisers Act of 1940. The investments and strategies discussed in TradeSmith's content do not constitute personalized investment advice. Any trading or investment decisions you take are in reliance on your own analysis and judgment and not in reliance on TradeSmith. There are risks inherent in investing and past investment performance is not indicative of future results.
Tidak ada komentar:
Posting Komentar