Tesla Inc. (TSLA, Financial) CEO Elon Musk has become notorious for setting extremely aggressive product development timetables. Such ambitious timelines have undoubtedly helped to buoy the electric vehicle maker’s high-flying share price on numerous occasions. Yet, in practice, Tesla has frequently struggled to stick to its boss’ timetable.
Indeed, Tesla has increasingly failed to bring Elon Musk’s promised products to market at all. Such has been the case with the Tesla Semi and next-generation Roadster, as I discussed recently. Both vehicles were unveiled to the public in 2017 and were supposed to enter commercial production years ago, yet both have ended up stuck in development.
While the company's various struggles to bring new vehicle models to market are notable, they arguably pale in comparison to its most notorious yet-to-exist offering: self-driving cars.
FSD dreaming
Musk has built a remarkable track record of claims about Tesla’s full self-driving technology, which the company has been selling since 2016. Unfortunately, while you can buy the FSD package, it cannot actually enable full autonomy, at least not yet.
When Tesla’s FSD technology will represent genuine self-driving remains an open question. Musk’s powers of prognostication have proven wanting thus far. As the following quotes compiled by investor and market commentator JC Oviedo attest, Musk’s FSD development timeline has been subject to frequent revision:
- October 2016: “Tesla expects to demonstrate self-driven cross-country trips next year.”
- January 2017: “[FSD will depart from Enhanced Autopilot feature in] 3 months maybe, 6 months definitely.”
- November 2018: “I think we’ll get to full self-driving next year.”
- April 2019: “We expect to be feature complete in self-driving this year...we expect to have the first operating robo taxis next year with no one in them, next year.”
- January 2020: “I was hoping we would be feature complete with both FSD by the end of last year. We got pretty close, it’s looking like we might be feature complete in a few months.”
- July 2020: “This is why I’m very confident about full self-driving functionality have been completed by the end of this year, because I’m literally driving it.
- January 2021: “I’m highly confident the car will be able to drive itself with reliability in excess of humans this year.”
Based on Musk’s statements alone, one might be forgiven for thinking the advent of truly autonomous vehicles is near at hand. Yet, as the above quotations attest, Tesla has had to walk back its FSD claims year after year.
AV anxiety
In 2015, Musk asserted that creating autonomous cars would be a relatively quick and simple task, claiming to “almost view it like a solved problem.” Yet, while Tesla has supposedly been on the cusp of releasing FSD ever since, the capability has remained remarkably elusive to this day. Indeed, Musk has lately commented about the unexpected challenges Tesla’s FSD development has faced. In April, he admitted that autonomy was proving to be “one of the hardest technical problems...that’s maybe ever existed.” Musk returned to this theme in comments last month:
“Generalized self-driving is a hard problem, as it requires solving a large part of real-world AI. Didn’t expect it to be so hard, but the difficulty is obvious in retrospect.”
The technical challenges posed by full autonomy may have only become obvious to Musk in retrospect, but they were already well understood by other AV developers. In fact, Musk has consistently been an outlier among industry leaders, most of whom have long admitted that achieving the highest level of autonomy, known as Level 5 in AV industry parlance, is still a decade or more away.
While Musk has begun to publicly admit to the true scale of the challenge, he has not backed down from his claim that FSD will be a reality in a matter of months.
AI Day daze
Musk has often waxed lyrical about Tesla’s capabilities in the artificial intelligence space, citing its focus on neural networks trained by real-world driving data as a critical advantage over other autonomous vehicle developers, all of which can access only a much smaller total volume of data, and many of which rely on simulations.
On Aug. 19, Tesla played host to its AI Day. While the event was billed as principally a recruiting drive for AI developers, it garnered significant investor and media attention as well. While Tesla presented some very complex and interesting information about the FSD development process, it also made some surprising admissions. Of particular note was the revelation that Tesla has turned increasingly to simulation to train the AI at the core of its FSD system. While this would be in keeping with the rest of the AV industry, it seems to contradict Tesla’s claimed advantage in real-world driving data.
Perhaps the strangest moment of AI Day came when Musk announced Tesla’s plans to develop a humanoid robot. To punctuate the announcement, Musk was joined on the stage by what appeared to be a person in a robot costume. This move raised many eyebrows, even drawing ridicule from a number of financial media outlets. While Musk claimed that Tesla’s “Bot” would soon revolutionize the economy, many industry commentators were doubtful. Automotive and AV industry journalist E.W. Niedermeyer, for example, offered a particularly scathing assessment:
“If you look at AVs and robotics more generally, the only viable products have been super focused: automate a very specific task and design it to fit the task's needs: Roomba, is one example. Welding cells in manufacturing. Specialized, not generalized. General tasks require general intelligence. We know how hard that is. We also know that humanoid robots have been tried. Development on Pepper the Softbank robot started in 2004, and it keeps getting fired from customer service jobs because it annoys more than it helps...Musk already tried an ambitious project in specialized robotics: ‘The Alien Dreadnought’ that was supposed to build cars with zero humans at blistering speed. That was such a disaster, Tesla now makes cars largely by hand in a tent.”
I have previously discussed Tesla’s ill-fated attempt to build a fully automated "Alien Dreadnought" factory, the failure of which led Musk to admit that “humans are underrated.” The company’s continued struggles with FSD can hardly be called a stirring endorsement of its ability to develop an AI-powered robot workforce anytime soon.
Profit imperilled
Tesla has been selling its FSD package since 2016. In the five years since, thousands of Tesla vehicle owners have spent thousands of dollars apiece on the FSD upgrade. This amounts to billions of dollars of what is essentially pure profit, since every Tesla vehicle is equipped with the same hardware. Those profits risk drying up if buyers begin to doubt whether Tesla can deliver on its promises. They could also be imperilled by government intervention. The National Highway and Transportation Safety Administration has opened dozens of investigations into FSD-related crashes in recent years, the latest of which could threaten Tesla’s controversial product, according to Slate’s David Zipper:
“So far, Tesla’s competitors have been considerably more cautious in crafting their own so-called Advanced Driving Assistance Systems. A smackdown from the feds could force Tesla to change course, while dissuading other automakers from emulating its risky behavior. But there’s another aspect of NHTSA’s investigation that looks promising: It focuses on a specific group of people who were outside the Tesla during a collision...These ‘vulnerable road users’ have no control over the design of American autos, which are growing heavier, taller, and more dangerous to people walking or biking. For years, the federal government has focused on automobile occupants, doing little to protect vulnerable road users beyond recommending that they ‘dress to be seen by drivers.’ But the investigation into Tesla offers a glimmer of hope that, at long last, that approach could be changing.”
Despite FSD revenue effectively going straight to its bottom line, Tesla has rarely managed to post an operating profit, instead relying on the sale of regulatory credits to keep it in the black. Absent FSD sales, Tesla would likely struggle to break even from its operations alone.
My take
While Musk’s many fans have largely shrugged off the repeated delays to date, I am skeptical of Tesla’s ability to keep kicking the proverbial can down the road without cost forever. Tesla says FSD will finally be ready for prime time this year. If it once again fails to deliver, the patience of Tesla owners and investors alike will once again be put to the test.
In my assessment, the longer Tesla delays FSD, the greater the danger that investors turn against the soaring stock.
