Musk’s robotaxi plan shows he gets cars but not people

By Dave Lee

Bloomberg News reported last week that Elon Musk — you won’t believe this — was going to miss a deadline. His make-or-break robotaxi unveiling would be postponed from August until October, a delay he confirmed in a post on 'X' on Monday.

When the news broke, shares in Uber Technologies Inc. and Lyft Inc. climbed sharply — Uber by more than 6 per cent, Lyft by around 4 per cent — suggesting some degree of relief from investors. There is a view on Wall Street that the introduction of Tesla Inc.’s self-driving fleet, now the company’s existential priority since ditching plans for a lower-cost car, spells bad news for the economics of the existing human-powered ride-hailing industry.

One analyst report described the two-month delay as pushing back the “imminent threat” to Uber’s and Lyft’s profit margins on the theory that Tesla’s large customer base in big cities means a vast self-driving fleet is ready for deployment, like a coiled spring. Others see small-scale rollouts of the robotaxi project beginning next year before becoming more broadly available in the first half of 2026. “This 2-month delay could just make the actual Robotaxi event and prototypes even better and more eye-popping for Tesla,” wrote Dan Ives of Wedbush.

It’s in times like this I wonder if we’re living in an alternate reality, one where the many broken promises, missed targets and obfuscations of Elon Musk never happened.

Uber and Lyft investors can relax. Musk, who began saying full self-driving was imminent in 2016, is still years away from his robotaxi ambitions coming close to fruition. I’m skeptical it will ever happen — at least not in the way he’s laid out and at a scale that would justify his pivot away from producing a lower-cost electric vehicle. Musk’s robotaxi strategy is a house of cards of deeply flawed assumptions.

The first is the belief that millions of Tesla owners will gladly send their own vehicles out to pick up members of the public in return for a share of the fare. To the Tesla owners reading this, I ask: Would you? Would you send out your car — which can’t have any of your belongings in it, unless you want them stolen — to go pick up the fast-food-eating, drunk-puking, sex-having general public?

The car would need to be thoroughly cleaned each day. Income for owners would most likely be paltry because the model only works if it undercuts Uber. Zero driver costs will help, but insurance will likely be high, not to mention any other associated fees cities will look to levy on a new mode of transportation. The times at which potential ride-share earnings are at their highest, such as at rush hour or for the school run, are likely the same times drivers will want the car for themselves.

Ultimately, if you’re wealthy enough to own a Tesla, I’m going to say it’s more likely you’d prefer to keep your car in good condition, constantly at your beck and call, and with a much slower rate of depreciation. The low single-digit dollar amounts per trip just won’t be worth it.

Now, Musk has suggested people (or businesses) could manage small fleets of Teslas — especially if, as some have predicted, Tesla’s robotaxi prototypes include a low-cost, low-frills self-drive-only vehicle. Tesla also says it will operate its own fleets. But, realistically, how many will invest in putting these cars on the roads when the margins are so small and the risk is so high? Where do they recharge? Who picks them up when they break down? Cost isn’t the only competitive moat the robotaxi must cross: There are 7 million Uber drivers globally. They handle sun, wind, snow and rain, with an average wait time of less than five minutes.

Another assumption is that cities will allow Musk to do what he wants. “I actually do not think that there will be significant regulatory barriers,” he said to investors in April, “provided there was conclusive data that the autonomous car is safer than a human-driven car.”

He’s wrong. The bar is higher. GM’s Cruise cars proved safer than the average human driver on the streets of San Francisco, but it took just one accident — the kind that sadly takes place all too frequently on American roads — to pull all of the company’s cars off the city’s streets (they haven’t returned). Safety isn’t the only question — much of the public’s frustration with current self-driving experiments is simply that the cars often get stuck.

Musk’s reputation will deservedly precede him. His lack of honesty around the actual capabilities of his “full self-driving” and “autopilot” technology to date has been shameful, prompting a recall and a government safety probe. City officials will be under immediate pressure from communities to not let self-driving Teslas anywhere near them — certainly not where they might be most profitable, such as near schools, busy shopping areas or streets teeming with nightlife. Parents will read articles about the teenager struck by a Tesla on “autopilot” driving at “highway speeds” and say: No, not around my kids.

While some companies, like Alphabet Inc.’s Waymo, have been able to launch impressive city-sized self-driving ride-share services, these have been tightly controlled and only made possible by more than a decade of delicate policy and safety lobbying. Musk lacks diplomacy and patience, which is sometimes a virtue, but not with this. Unable to contain his frustration, he’ll no doubt jump on to X in the wee hours one morning and lash out at those standing in his way. It won’t help. Local politicians may fear a Musk tweetstorm but not nearly as much as they fear the angry parents at the school gate.

And then there’s the question of how many people will be willing to ride in one. I’ve used Tesla’s “full self-driving” mode while sitting in the driver’s seat several times and marveled at its incredible — though limited and unpredictable — abilities. The robotaxi will be more sophisticated, one would hope, but Musk has so far remained stubborn in his highly criticized view that cameras are enough to provide safety. As it stands, I’d no sooner want to ride in a self-driving Tesla through a busy street than I would on a spooked elephant.

Can Tesla create a safe, genuinely fully self-driving car? It’s possible. Musk has shown that he excels at managing teams that solve engineering challenges that others previously thought were nearly impossible. But in some respects, the car isn’t truly the challenge. Musk has demonstrated a deep-rooted inability to understand human behavior. The mass adoption of self-driving vehicles is primarily a societal question. We have deep-rooted fears and concerns about safety and job displacement. Other companies are better placed to manage this, and Uber is one of them. Its position as a wholesaler of both human drivers and self-driving cars, as telegraphed by Chief Executive Officer Dara Khosrowshahi, is a far bigger threat to Tesla’s ambition than Tesla is to Uber’s.

Musk is ignoring these complexities, and investors have been too willing to let him get away with it despite the warnings hiding in plain sight. “If somebody doesn’t believe Tesla is going to solve autonomy,” Musk said in April, “I think they should not be an investor in the company.” Sound advice, Elon.