It's fair to say that Tesla's 'We, Robot' event was a totalcar crash, and barely even metaphorically. It was more akin to FSD fatally steering a Tesla into a bridge at highway speeds — a totally avoidable mess caused by misinformed people with a massive overreach of reckless power. The event was so bad that Tesla's shares have lost a staggering $60 billion in value. And that is just the beginning. You see, the 'We, Robot' event uncovered the deadly rot at the core of the American giant. Tesla is dying, and here is why.
Before we dive into the shitstorm that was 'We, Robot,' we must understand the context for how important this event was. Musk forced Tesla to sacrifice pretty much everything that gave the company a market advantage to bring fully autonomous cars and robotaxis to the market. The fast charging network that made Teslas the first usable EVs, gigacasting development that made Teslas affordable, and even new vehicle development crucial to fleshing out Tesla's sparse and ageing lineup have been wholly dispanded or outright demolished to save money and pump more funds into Tesla's self-driving AI. This is nothing new; Musk has prioritised the AI drive over crucial developments such as Tesla's 4680 battery, which was meant to give them a considerable leg-up but has mostly been a complete letdown thanks to slow development caused by this redirection of funds.
This has enabled the competition to catch up or even overtake Tesla in terms of price, specs, and profitability. As such, Tesla's EV market share is tumbling and is only set to fall further. So, in order for Tesla to keep its wildly overvalued market cap and position in popular culture (which is vital to Musk's brand and ability to raise money), it has to bring fully autonomous self-driving cars and robotaxis to market ASAP.
This event was a chance for Tesla to show that they can deliver on this promise — not just show off some cruddy concept vehicle with technologies that don't yet exist and provide unfounded placation as to when this will come to market like Musk has done time and time again for well over a decade by now. So, did Musk take this chance to prove the naysayers wrong and demonstrate some actual substance?
Well, what do you think?
There was nothing of significance in the event. It was a glitzy PR stunt, with nothing to back up their business model, technology, and roadmap or make them stand out against the sea of other autonomous vehicle companies that already have robotaxis in service.
During the event, Musk unveiled their "Cybercab," a two-seater steering wheelless and pedalless robotaxi that looks like a cross between the Cybertruck and VW's brilliant XL1 car. Musk also unveiled a Robovan that looked like it was pulled straight from an Art Deco dystopian future and had a small army of Tesla robots (also known as Optimus robots) serving guests drinks. During the event, Musk explained that Tesla will bring fully autonomous driving (also known as "unsupervised" Full Self-Driving) to Tesla in Texas and California in 2026. He also stated that the Cybercab will cost "less than $30,000" and that they will start mass producing it before 2027.
These are some huge claims, so what evidence did Musk have to back up these plans? Does Tesla have explanations of how their technology will work? Data to evidence that they can actually deliver a safe self-driving vehicle? What about robust demonstrations of said technology?
Nope. None of that. There were a handful of concept robotaxis driving attendees around. Still, they were driving slowly around the private Warner Bros. Discovery Inc.'s movie studio where the event happened, a far cry from operating on actual roads. I could program a car to do what these concept vehicles did, and I can barely build a program in Python. Musk hardly talked about the technology behind the Cybercab and gave no data on how Tesla's FSD self-driving AI is getting better or would get better. In short, nothing about the event suggests that Musk could even come close to delivering on these promises.
And it gets worse. This event highlighted problems with the Cybercab, FSD, and Tesla's Optimus robots, which demonstrate that Tesla can't deliver on these promises.
Let's start with the Cybercab. Firstly, experts have damned the design, criticising it for only having two seats and a tiny interior space, making it practically useless for a lot of taxi jobs, like taking families to airports. They have also ripped into it for its low ride height, which will make accessibility impossible for anyone with mobility issues and reduce the effectiveness of the self-driving sensors, given that because they are lower to the ground they will get a worse view of the world around them. This is why other robotaxi companies like Waymo use tall, large vehicles, as it helps with every aspect of their design brief. In short, the Cybercab design makes for a terrible self-driving taxi.
But it also has some seriously fatal flaws in its self-driving technology.
You see, the Cybercab still uses Tesla's vision-only sensor technology. This means it doesn't have the lidar, ultrasonic, or radar sensors that almost every other self-driving company uses and instead only relies on camera feeds to understand the world around them. Now, there is a reason these companies use these extra sensors: it gives their system redundancy. Camera feeds can fail due to bad lighting conditions, ambiguous visual data, or failure of the computer vision AI, which translates the camera feed into a 3D map. These other sensors aren't infallible; they just fail in different ways, but they do give far more consistent 3D data than cameras. As such, by running multiple types of sensors at once, you can use them to cross-reference each other and remove anomalous readings or AI driving outputs that would otherwise be dangerous. This, in turn, means the AI driving the car doesn't have to be so accurate, as it has a safety net around it, and so it is easier and cheaper to build. Meanwhile, as Tesla only uses cameras and their cars have very few, their AI has to be extremely accurate to even hope to pass safety regulations.
But as I have covered before (read here), we are seeing woefully diminishing returns on AI training, and it is beginning to look like you physically can't mitigate AI errors, no matter how much data you feed into them. As such, Tesla can't make an AI accurate enough to make their vision-only approach work. In fact, Tesla's own engineers warned Musk of this when they switched to it, and they have even whistleblown to the papers claiming Musk's move to vision-only makes the system incredibly dangerous.
These claims were backed up by data, as since then, there have been a sizeable number of deadly accidents caused by Tesla's FSD. There have been so many of these accidents that Tesla is under investigation by several government authorities and is facing a mountain of lawsuits.
With this in mind, it is no wonder Tesla has been very secretive about the safety data of its FSD software (which its Cybercab software will be based on). As such, we don't know how dangerous the system actually is.
However, a recent third-party investigation found that, on average, Teslas driving "autonomously" with the latest version of FSD had to be intervened by the driver every 13 miles. It's important to note that drivers didn't intervene every time FSD did something wrong, just every time it did something massively dangerous, like driving at a cyclist, failing to recognise stop signs, or encroaching onto oncoming traffic. So this figure is utterly damning!
We can also use this data to determine how far Tesla is from producing a robotaxi that is as safe as a human-driving taxi. Let's be extremely generous and assume that only 1% of these interventions would have led to a crash. This means that in its current guise, FSD can travel around 1,300 miles on average before it has a crash. The average taxi drives around 40,000 miles per year. As a result, if the Cybercab used today's version of FSD, each Cybercab would get into an average of 30.7 crashes a year. To give you a sense of context here, a human-driven taxi gets into an average of 0.04 crashes per year.
In other words, even with an incredibly generous assumption of how bad these interventions were and assuming the Cybercabs would work the same hours as a regular taxi driver, not more, they would still be roughly 767.5 times more dangerous than a regular taxi!
This is why the fact that Musk didn't detail any safety data at 'We, Robot' was a knell in the coffin. This has been a glaring issue for Tesla for years, with even the most ardent Tesla bulls questioning the safety of Musk's vision-only FSD. Musk would have known that if he could back up how safe FSD is, it would make the Cybercab pitch fly, and without it, no one would believe him. But he couldn't do that; it would be too much of a lie, even for Musk.
Speaking of lying, it seems Musk resorted to deception at this event yet again. Those Tesla robots pouring drinks were meant to demonstrate how capable Tesla's automation is, reassuring attendees and the media that Tesla can build dependable AI automated systems. However, it seems these robots weren't automated but puppeteered by employees behind the scenes. Musk isn't a stranger to this sort of sleight-of-hand marketing; back in 2016, he orchestrated a fake self-driving press video to boost the company's value and drum up investment.
But, getting back to the point, this lack of safety has made many question whether Tesla can meet its targets. Even though Texas and California have looser self-driving laws than other states or countries, there are still rigorous safety tests and standards Tesla needs to meet to offer eyes-off-the-road self-driving in the Model 3 and Y. The current third-party safety data and horrifying lack of safety data from Tesla show that it simply can't meet these standards or pass these tests, even in the most relaxed regulatory settings. Worse still, the regulations around offering a vehicle with no pedals or steering wheel, like the Cybercab, are exponentially tighter, as you need to guarantee that the driver will never have to intervene. As such, all the evidence points to Tesla not being able to sell the Cybercab by 2027 or even near that deadline.
Now, everything we have already covered is enough to damage Tesla severely. But believe it or not, it gets worse! You see, Tesla has the worst possible business model for rolling out their robotaxis.
All other robotaxi companies plan to be owner-operators. They will develop, own, and operate the vehicles. This makes responsibility and liability crystal clear. If there is an accident, they are the ones liable. Likewise, it's their fault if there is bad service, like a dirty car. This streamlines the legal process, enabling them to scale up production faster. It also means they have control over their service and can be held responsible for bad service. This encourages cities to work with and welcome them, as they can work together to make the robotaxi rollout and operation beneficial and successful.
But Musk wants to sell the Cybercab to people who will then rent it out as a robotaxi. Hence, he proclaimed it would "cost less than $30,000." Such a price doesn't matter to other self-driving companies, as they can still turn a profit even if it costs double that.
You might think that Musk's approach is smart, as it offloads the burden of scaling a robotaxi business onto others, allowing them to focus on manufacturing them. But that makes no sense. Tesla has a rich history of dodging liability for accidents that involve its self-driving software, so who will be responsible when one of these Cybercabs inevitably crashes? Will potential buyers have a guarantee that they won't be held liable? How can Tesla ensure a good service? Most potential owners don't have time to clean these vehicles between trips, and those that do might chose not to in order to maximise profits. With this in mind, why would a state country or city allow Cybercabs on their street if they are legally dubious and they can't work with them to ensure the service actually benefits the citizens?
Moreover, much of the hype around Tesla's robotaxi push has been for Tesla to be an owner-operator. Cathie Woods has repeatedly said that Tesla's robotaxi could secure Musk billions upon billions of dollars in revenue from annual fares. But if Tesla has no plans to actually own and operate them themselves, this can't happen!
In short, the business model Musk is pushing for the Cybercab is terrible and completely undermines many of Tesla's most prominent investors.
This is why Tesla's value dropped by $60 billion after 'We, Robot.' It wasn't just the Tesla sceptics like me that watched in disbelief. It was the hoards of Tesla bulls, retail investors, and institutional investors alike, watching as Musk demonstrated how Tesla can't deliver what they hoped. This event showed them that Tesla's fundamentals and reality don't align with the hype and, in doing so, kickstarted a mass sell-off. Even historic Tesla backers like Morgan Stanley expressed disappointment and sold off. In fact, many analysts who are Tesla backers have said that this drop in value is likely to get far worse as more and more people sell up over the coming months or large institutions stagger their sales to avoid crashing its value too quickly.
Okay, so how is this the death of Tesla? Surely, they still have their EV branch to fall back on.
Well, possibly not for long.
Bernstein's Toni Sacconaghi has valued Tesla's automotive business at $200 billion, meaning $600 billion of its current value relies on a successful shift to robotaxis. 'We, Robot' demonstrated that Tesla doesn't have the substance to see this shift through even remotely successfully, meaning it is far from being worth $600 billion. Not only that, but the fact they have sacrificed and neglected almost everything that made their automotive side of the business and that their market share is rapidly shrinking suggests that that side of the company might not be worth $200 billion for long.
Furthermore, all the talent that made Tesla so dominant within the EV market upped and left in the week following 'We, Robot.' Tesla's CIO, public policy chief, Model X program manager, and the head of vehicle programs all left, citing abuse by Musk as their reason for leaving. These people were crucial to Tesla's past success, most notably the head of vehicle programs, Daniel Ho, who had been at Tesla for ten and a half years and oversaw the successful rollout and scaling of the Model 3 and Y. The situation inside Tesla is so bad that one former manager said executives were left 'just fighting to keep our teams together.' 'That s*** takes its toll.'' 'Every few years Elon comes in and slashes headcount or there's a reorg, and it's like you have to build everything from scratch again.' 'People get burned out.'
With all of Tesla's funds focused on making the Cybercab a reality and Tesla haemorrhaging crucial talent at an alarming rate, they likely can't hold onto their position in the EV market. As such, if Tesla can't get the Cybercab off the ground and selling like hotcakes, there is a question of whether they will remain profitable and sustainable or if they will face a crippling squeeze.
Because the Cybercab is likely a lost cause, it seems Elon Musk has checkmated himself, and Tesla will pay the price for it.
However, there is another reason why Tesla is dying.
Musk's record-breaking pay packet was insane. It was the equivalent of him receiving $10,000 for every car Tesla ever sold! Tesla actually makes less profit than that for many of their models, but that doesn't matter to Musk. He uses Tesla as a way to manipulate the stock market. Historically, he has been able to inflate or deflate the value of Tesla on a whim to suit his purposes by faking technological breakthroughs, driving hype through the media, or joking that the company has gone bust. This has enabled him to sell, buy, or leverage his stocks efficiently, or drum up investment or use these stocks as collateral at any time. This has kept him engaged with Tesla and made him a vast amount of money, as it has enabled him to grow SpaceX or buy Twitter. However, 'We, Robot' was the first time he tried to manipulate the stock market this way, and it has backfired. Even institutional investors who happily rode his coattails before by peddling his lies have turned on him. Musk now knows that investors will only back Tesla if they demonstrate a clear ability to deliver on his hollow promises, which they can't do.
As such, I predict Musk will become less and less interested in Tesla as it doesn't suit his needs any more. He will stop driving its hype and instead keep cutting costs to keep it buoyant, and eventually, Tesla will eat itself alive from the inside out.
That is how the 'We, Robot' event kickstarted the death of an American Giant.
Thanks for reading. Content like this doesn't happen without your support. If you want to support content like this, or read articles early, go and follow me and my project Planet Earth & Beyond.
(Originally published on PlanetEarthAndBeyond.co)
Sources: Yahoo, CarScoops, Benzinga, AOL, AOL, The Verge, Gizmodo, Reuters, BBC, The Guardian, Autocar, Daily Mail, Flying Penguin, Electrek, Cleantecnica, Interesting Engineering, Taxi Point, Nimblefins