Tesla’s robotaxi network, still in the implementation phase, traveled a shorter distance to paying customers in the second quarter than in the first, as indicated by a chart released by the company on Wednesday, August 12, 2026.

This quarterly reduction contradicts Tesla’s strategy and statements over the last year. The company has invested significantly in the concept of a vast fleet of robotaxis, aiming for low cost and high revenue generation, a plan that CEO Elon Musk described in 2024 as “investing everything in autonomy”. The decrease in robotaxi mileage also coincides with a drop in profits from Tesla’s core business, which fell below Wall Street projections, according to data also revealed on Wednesday. Tesla shares fell more than 13% at the start of trading on August 13, 2026.

While the graph, at first glance, suggests continued progress in paid robotaxi trips between August 2025 and June 2026, a quarterly analysis of the cumulative numbers reveals a different story. Tesla’s robotaxi fleet, made up of Model Y SUVs that carry passengers, logged about 1.1 million miles in the first quarter. In the second quarter, that number dropped to approximately 700,000 miles, indicating a decline of about 36%.

The decrease occurred even as the company expanded its initial operations to six cities in Texas and Florida, using a combination of human-supervised and unsupervised vehicles.

It is likely that Tesla is including paid kilometers in the San Francisco Bay region in the count. However, these customized robotaxis do not have the necessary state licenses to operate autonomously and still depend on the presence of a safety driver at the wheel, with this operation referred to by Tesla as part of its “Robotaxi coverage”.

The decrease in mileage also coincides with a revelation made by Tesla during a Q2 earnings call on Wednesday, August 12, 2026. Asked about the slow expansion of the robotaxi service, Elon Musk said the company needs to “accumulate specific driving data for the Cybercab” — the two-seat sedan designed to be the basis of the future autonomous fleet — “before we can put many of them on the streets.”

Musk explained that, unlike models like Model 3 and Model Y, which already have millions of units running, Cybercab does not have this vast amount of data. Therefore, it is necessary to collect mileage with adapted Cybercabs, equipped with steering wheels, accelerator and brake pedals, to calibrate them with the specific chassis. He expressed confidence that once this is done, the presence of Cybercabs in cities will increase significantly.

This statement marks a shift from the company’s previous claims, which for years maintained that its fleet of nearly 10 million customer vehicles was continually collecting data in the background to train future robotaxis and the driver assistance software known as Full Self-Driving.

During the call, Tesla directors justified the slow pace of development as a safety precaution.

“Our goals for the Robotaxi are very ambitious, but we need to be careful not to cause accidents or harm anyone,” said Musk.

He added that he does not want Tesla robotaxis to be involved in accidents, as he believes that negative press coverage could result in greater regulatory rigor.

“While, I believe, 30,000 to 40,000 car accident deaths occur annually in the United States alone, most of them do not generate press reports. However, if we injure a single person, it will make global headlines, and regulators will immediately restrict our activities,” the executive said.

Ashok Elluswamy, vice president of AI at Tesla, highlighted that the company’s robotaxis have had “no notable incidents” while traveling “more than 380,000 miles” without a safety operator on board. He did not detail what the company classifies as “notable incidents”, but clarified that “any reports refer to other agents hitting us when we were stopped.”

Tesla has reported 22 accidents to the National Highway Traffic Safety Administration (NHTSA) since testing its robotaxi service began. While most involve other vehicles colliding with Tesla robotaxis, the company has recorded three accidents caused by its remote operators and several cases of vehicles hitting objects at low speeds, such as curbs, light poles and the body of a tow truck.

This situation constitutes yet another change in the company’s line of communication. For years, Tesla has claimed that the biggest impediment to the large-scale deployment of robotaxis was regulatory in nature, although it never specified what those restrictive regulations would be.

Currently, the company argues that proving safety is the only obstacle to the advancement of robotaxis. Even in its early stages, Tesla took the opportunity to reinforce its decision to develop an autonomous driving platform that does not rely on sensors such as radar or lidar, unlike industry leaders like Waymo.

“Historically, so-called experts have always argued that you need lidars, radars, HD maps and everything else to drive safely. Here, we demonstrate that this is not true. You can have safe, comfortable and affordable autonomy with just cameras,” said Elluswamy.

Both Musk and Elluswamy made promises of growth, noting that unsupervised mileage driven has increased by about 10% per week since Tesla began offering the service late last year.

“I believe we will continue to grow very, very quickly,” Musk said.