Tesla’s emerging “robotaxis” network drove fewer miles for paying customers in the second quarter than in the first quarter, according to a chart released by the company on Wednesday.
The quarter-to-quarter decline stands in contrast to Tesla’s rhetoric and actions over the past year. Tesla has staked much of its future on the idea of a huge, low-cost, cash-generating robotaxi fleet — or the idea of putting “balls to the wall for autonomy,” as CEO Elon Musk framed it, in 2024. The quarterly decline in robotaxi miles also comes amid weak profits at Tesla’s core businesses, which have performed below Wall Street expectations, according to data released Wednesday. Tesla shares fell more than 13% in early trading on Thursday.
At a glance, the chart shows a steady increase in paid robotaxi rides between August 2025 and June 2026. But the numbers displayed are cumulative, and when broken down by quarter, they show that Tesla’s robotaxi fleet of Model Y SUVs drove paying passengers about 1.1 million miles in the first quarter. That dropped to about 700,000 miles in the second quarter, a decline of about 36%.
This is despite the fact that the company has expanded its initial operation to six cities in Texas and Florida with a mix of unsupervised and supervised vehicles.
It’s possible that Tesla is also counting paid miles driven in the San Francisco Bay Area, even though these branded robotaxis do not have state-required permits to operate autonomously and even have a safety driver behind the wheel. Tesla refers to that operation as part of its “robotaxis coverage.”
The decline in miles driven also comes as Tesla made an important admission regarding its second-quarter results in a conference call on Wednesday. In response to a question about how slowly Tesla is scaling up the robotaxi service, Musk said the company “needs to accumulate specific driving data for the CyberCab” — the company’s gilded, purpose-built, two-seat sedans that are expected to form the bulk of its autonomous vehicle fleet — “before we can put a lot of them on the road.”
“Unlike the Model 3, Model “As we become convinced of this, the number of cybercabs in cities will increase dramatically.”
This is somewhat of a departure from claims the company has made for years about how its fleet of nearly 10 million customer cars is quietly collecting data in the background to train future robotaxis (in addition to training driver assistance software for consumers, what Tesla calls full self-driving).
On the call, Tesla executives described the slow progress as a matter of being cautious about safety.
“Our goals for robotaxi are very ambitious, but we need to be cautious about any accidents or causing any harm to anyone,” Musk said.
He then said that he did not want Tesla robotaxis to cause accidents because he felt that poor media coverage could lead to regulatory action.
“Although there are, I think, 30 to 40,000 motor vehicle deaths per year in the United States alone, most of them generate no press, you really don’t read about almost any of them. But if we injured even one person, it would be headline news around the world, and regulators would immediately crack down on our activities,” he said.
Ashok Eluswamy, Tesla’s vice president of AI, claimed that Tesla’s robotaxis have had “zero notable incidents” while driving “over 380,000 miles” without a safety operator. He did not define what the company considers “notable incidents”, although he claimed that “there have been no reports of other actors influencing us while we were stable.”
Tesla has reported 22 crashes to the National Highway Safety Administration in the year since it began testing its robotaxi service. While most of them involve Tesla’s robotaxis crashing into other cars, the company has reported three crashes caused by its teleoperators remotely piloting the vehicles, and several instances of cars hitting objects at low speeds, including curbs, utility poles and the bed of a tow truck.
This represents another narrative change for the company. Tesla spent years claiming that the biggest hurdle to full-scale robotaxi deployment was regulatory in nature — though the company never actually specified what those prohibitive regulations were.
Now the company says that proving safety is what is stopping Robotaxis. And although it is still in the very early stages, Tesla chose this moment to trumpet its decision to create an autonomous stack that does not use radar or lidar sensors like the industry leader Waymo.
“Historically, so-called experts have always claimed that to drive safely you need lidar, radar, HD maps, and the entire kitchen sink. Here, we show that’s not true. You can have safe, comfortable, and affordable autonomy with just cameras,” Eluswamy said.
Both Musk and Eluswamy also promised that development is coming. He noted that unsupervised miles traveled have increased by about 10% each week since Tesla began offering it late last year.
“I think we’re going to continue to move very quickly,” Musk said.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
