Tesla's Austin robotaxis go driverless, and nobody will confirm it

Seven months after Elon Musk claimed Tesla's Austin robotaxis were running without human safety monitors, a crowdsourced tracker says the city's service has finally gone fully driverless. The catch: Tesla still won't say so itself.

What actually happened

Over the past two weeks, every one of the 170 Tesla Robotaxi rides in Austin tracked by the volunteer-run Robotaxi Tracker ran with no human safety monitor onboard, the site's creator, Ethan McKanna, told The Verge. Those rides spanned 54 different cars. McKanna also reports roughly 30 unsupervised Teslas operating in Dallas and Houston over the past week — a sharp reversal from earlier this summer, when his own tracker showed the unsupervised fleet shrinking, with just 28 active driverless vehicles across six Texas and Florida markets as recently as last week.

The Austin number alone is nearly double what his tracker had previously logged across all six markets combined, which sounds impossible until you account for data quality. McKanna, who spent this summer interning with Tesla's Robotaxi team, says some metrics on his site were lagging and undercounted real activity; he is revising the sources feeding the tracker. But he insists the Austin shift is real, not a measurement artifact: plates previously flagged as supervised have been confirmed offering unsupervised rides, and rider reports on Reddit and X back him up.

That last point matters. Tesla does not publish fleet data and did not respond to requests for comment. So the strongest public evidence that the company has pulled humans out of its cars comes from a college student's spreadsheet and a handful of passenger anecdotes — not from Tesla.

Why the timing is deliberate

This is not a coincidence of operational maturity. Tesla is preparing to launch the Cybercab, a purpose-built robotaxi with no steering wheel and no pedals, in Austin as soon as this month. A vehicle with no controls only works if the company can legally and technically run it without an onboard monitor. Going unsupervised in the Model Y fleet now is the prerequisite rehearsal for the Cybercab rollout, and it explains a second quiet move: according to a recent JPMorgan note, Tesla is intentionally holding back on adding more Model Y robotaxis because it expects to scale Cybercab instead.

The strategic logic is coherent. Tesla's whole bet is that a vision-only, camera-based autonomy stack trained on consumer fleet data can reach driverless capability far cheaper than the lidar-and-high-definition-maps approach its rivals use. If it works, Tesla's marginal cost per robotaxi is a retrofitted Model Y or a stripped-down Cybercab — not a bespoke, sensor-laden vehicle. Removing the safety monitor is the moment that bet stops being theoretical.

Who's winning, who's not

Read against the rest of the field, Tesla is simultaneously ahead and far behind. As we covered in Tesla puts ByteDance's Doubao AI inside its cars, the company is already weaving AI into the cabin experience; autonomy is the harder, higher-stakes layer. And on raw scale, Tesla trails its main rival badly. Waymo runs more than 300 driverless vehicles in Austin alone and roughly 4,000 across eleven cities — the kind of footprint Tesla's tracker suggests it is only beginning to approach in a single market. Our earlier look at Waymo's fleet hits 4,000 as recalls and edge cases mount showed that scale brings its own headaches, from parking fines to regulator scrutiny.

Tesla's advantage is different: speed of iteration and a manufactured fleet. Waymo expands city by city under bespoke permits; Tesla flips a software switch across cars it already built. That is why Nevada's decision to cap Tesla's Las Vegas robotaxi fleet at 10 vehicles is a tell — regulators are treating Tesla's unsupervised ambition as a live risk, not a future one.

The skeptic's case

The contrarian read is that "fully unsupervised" is doing a lot of work here. Robotaxi Tracker is a crowd-sourced project, not an audit. McKanna himself is a Tesla insider by virtue of his summer internship, which is exactly the kind of proximity that can bias both the data and the interpretation. Self-reported rider posts on Reddit and X are anecdotes, not evidence of a systemic, city-wide policy change. And Tesla's own silence is the loudest signal: a company that wanted credit for a genuine milestone would say so on an earnings call. Instead, Musk offered only that the program would scale "very, very rapidly," with no new vehicle or city targets — a retreat from the specific forecasts he has repeatedly missed.

There is also the safety framing. Going driverless in Austin with a vision-only stack, before a purpose-built vehicle exists, means the early unsupervised miles are running on the same Model Y hardware that has drawn regulatory and recall attention elsewhere. The absence of a human monitor is precisely when edge cases become incidents.

What to watch next

Three things will tell us whether this is a real inflection or another over-read of thin data. First, does Tesla publicly confirm the unsupervised status before the Cybercab launch, or keep letting trackers do the reporting? Second, does a regulator — Texas, Nevada, or the NHTSA — step in once "unsupervised" is treated as confirmed rather than reported? And third, the Cybercab itself: a vehicle with no controls is an all-in bet on the same autonomy that Austin is now quietly testing. If the launch slips or ships with monitors after all, the driverless Model Ys look less like a milestone and more like a pilot.

Should robotaxi operators have to disclose their supervision status, or is silence just part of the race? Tell us in the comments.

Sources: The Verge · Robotaxi Tracker · Tesla Cybercab launch coverage · JPMorgan note via Electric Vehicles · WSJ on Waymo's Austin fleet