The Tesla Cybercab has now entered series production. Yet the vehicles might just as well be shipped straight to the scrapyard. They do not work. Tesla’s CEO hopes they soon will — and that hope is precisely why he is building them.

How do we know Cybercab’s autonomy doesn’t work?

Because it relies on the exact same underlying system as every other Tesla, and that system does not deliver true self-driving capability.

Removing the wheel and pedals will not suddenly make the technology function. That is magical thinking, not engineering.

The Teslas currently operating in Tesla’s experimental taxi service in Austin are driving under supervision. There is no compelling reason to believe they possess the robustness required for fully unsupervised operation.

It is not the Cybercab roaming Austin, but ordinary Model Y vehicles running essentially the same software used across Tesla’s fleet for supervised driving. Every one of them retains a steering wheel, pedals, and a safety driver.

Tesla has never dared to let its Austin taxis operate without oversight. At the very least, no credible reports suggest otherwise — and Tesla’s trials are scrutinized relentlessly. Private observers even log which vehicles, down to license plates, are being deployed.

From time to time Elon Musk attempts to persuade the public that Tesla’s cars drive themselves. Those claims have invariably unraveled. In Austin, the human supervisor is sometimes positioned in a trailing vehicle, which briefly fueled public excitement after one of Musk’s characteristically cryptic tweets.

Despite the fact that these vehicles operate under supervision — and presumably handle only the simplest driving scenarios — some statistics suggest they are involved in accidents more frequently than typical human drivers. The numbers must be treated cautiously: the fleet is small, crashes are rare events, and the driving patterns are far from comparable. The most recent figure I encountered claimed four times as often, though I have not verified it.

Note that we are still discussing a single location: Austin, Texas — a city with forgiving weather conditions.

Nothing indicates Tesla possesses technology anywhere near capable of genuine unsupervised autonomy, even there.

The most likely outcome, then, is that discussion of Cybercab production gradually fades in the coming years — much as it did with the Tesla Semi.

Tesla Semi’s mass production slipped year after year, accumulating roughly six years of delays. Musk has already signaled that scaling Cybercab production will also take a very long time.

Waymo, by contrast, has cautiously expanded its unsupervised robotaxi operations over seven years since its debut in late 2019. Even so, it has yet to venture into truly challenging climates, and it limits service to carefully selected urban zones.

Even if Tesla’s autonomy were to prove itself in Austin a few years from now, every other city would remain an unsolved problem.

My expectation is that Cybercab will never become a high-volume product — not with Tesla’s current camera-only sensing strategy.

Perhaps it may find narrow niches: tightly defined routes in benign environments. There are obvious low-hanging opportunities, like transport distances between airports.

There is talk of Teslas driving themselves out of the factory. That, in practice, represents the furthest Tesla has genuinely advanced its autonomy: from the assembly line to a nearby parking lot.

Tesla’s steering-wheel-less robotaxi remains, for now, exactly what it has long been — a promise.

And it is a promise grounded not in demonstrated capability but in the wishful thinking of Tesla’s CEO, Elon Musk.