When Elon Musk kicked off Tesla Inc.’s quarterly earnings call in July 2025, robotaxis were top of mind. The carmaker had just launched its automated ride-hailing service several weeks earlier in Austin, Texas, to glowing early reviews from the hand-picked crop of initial riders.
“We’ve already expanded our service area in Austin,” the chief executive said. Tesla was planning to grow further in the city and quickly spread to California, Nevada, Arizona and Florida. “We’ll probably have autonomous ride-hailing in about half the population of the U.S. by the end of the year.”
Almost a year later, Tesla has just 59 vehicles in its entire robotaxi fleet as of Tuesday, limited to three Texas cities.
Musk is known for making overly ambitious proclamations and missing self-imposed deadlines at Tesla and SpaceX, the latter of which is literally in the business of moonshots. Yet even by those standards, the gulf between his public robotaxi statements and what the business has achieved is vast. Three months after that earnings call, Musk said Tesla would have “500 or more” vehicles in the Austin area alone by the end of 2025.
The diminutive fleet figure, which Tesla hadn’t disclosed until new regulations made it public last month, is a sobering reminder of the real-world limitations of Musk’s hype machine. At a time when Tesla’s carmaking business is mired in a multiyear decline, the company’s market value has soared to new highs almost exclusively on the billionaire CEO’s fantastical promises of a future in which all cars will drive themselves and robots will babysit the kids. It’s all right around the corner, he says.
Yet there are few signs that any of those cutting-edge business ideas will contribute meaningfully to Tesla’s bottom line in the near term — or ever. The Optimus humanoid robot program remains in prototype. More recently, Musk announced that he wants to build a massive semiconductor factory that could take years to get off the ground and cost as much as $119 billion.
Robotaxis arguably represent Tesla’s most concrete source of new growth. But as the service approaches its one-year anniversary, the results so far have been disappointing. Wait times can be long, availability is limited and the cars pick up and drop off at incorrect locations, with human safety monitors occasionally riding along to supervise. Musk has already said the business won’t be profitable until at least 2027.
A few dozen vehicles after almost one year is “well short of even our very bearish outlook,” said Garrett Nelson, a senior equity research analyst with CFRA. “For well over a decade, Musk has overpromised. … Add the robotaxi rollout to the list.”
Part of the delay is the “extremely difficult technological challenge” that comes with launching a driverless ride-hailing service, Nelson said. To meet tough safety standards, Tesla is proceeding cautiously with its automated technology.
Indeed, the repercussions of any misstep can be severe. Cruise, the autonomous-vehicle business owned by General Motors Co., struck and injured a pedestrian in California in 2023 — an incident that led to state and federal investigations, the suspension of Cruise’s operating permit and, eventually, the robotaxi business being shut down.
“The potential financial liability and regulatory consequences of not getting it right are enormous,” Nelson said.
Tesla has employees in Austin and the San Francisco Bay Area who have the ability to remotely support and operate the robotaxis. Since its launch, the company has reported 17 incidents to the National Highway Traffic Safety Administration, including two where remote employees crashed the vehicle at low speeds.
Rivals have ramped up more steadily. Waymo, in particular, has vastly outpaced Tesla’s scale in its home state, with the Texas DMV website showing more than 600 automated vehicles registered by the Alphabet Inc.-owned business, or roughly 10 times Tesla’s total.
Robotaxi user experience varies widely by ride and day. While testing the service periodically in Austin in recent weeks, a Bloomberg reporter frequently encountered long wait times of up to 30 minutes and periods where the app showed “high service demand” and barred rides from being booked. Three out of 10 rides had human safety monitors in the front seat.
One driverless ride, scheduled from a Tesla showroom, failed to start after arriving. The car screen displayed red bars and an error message that read, “Please contact support for assistance.” The remote assistant advised hailing a new ride, which was 13 minutes away.
Pickups and dropoffs were at times unpredictable, with cars stopping across the street or several buildings away despite available parking out front. One ride to a coffee shop ended with a dropoff in an alley instead of on the street, temporarily blocking a truck making its way through the narrow lane. When the same vehicle was hailed from the coffee shop a short while later, it parked out of view on a different street instead of arriving at the designated address.
Those issues reflected similar ones shared by robotaxi users on social media. In January, shortly after Tesla started rolling out unsupervised rides, Tesla enthusiast David Moss posted on X that it took him 58 robotaxis before he was able to land an unsupervised one. Three months later, another rider said on X that his robotaxi dropped him off a quarter-mile from where it was supposed to, despite having a human monitor in the car.
“Failed Robotaxi ride,” the post said. “On our walk to our destination, we witnessed a Waymo dropping off a rider in the place that we should have been dropped off.”
Investor and longtime Musk supporter Cathie Wood rode in a robotaxi recently and described the experience as “remarkable” — except for one wrinkle.
“Wow. We got a $75 ticket,” Wood, the founder and CEO of ARK Invest, said in a video posted on her social media channels Monday. “Well, Tesla did.”
Musk has proved skeptics wrong plenty of times before, successfully making electric vehicles mainstream and revolutionizing space launches with reusable rockets. His similarly lofty promises for data centers in orbit and human colonies on Mars are a big part of why SpaceX is expected to stage the biggest IPO in history. Fans and investors keep buying in.
In April, Musk said wider robotaxi deployment was limited by “convenience issues” where the car gets “paranoid and gets stuck” because it’s “programmed for maximum safety.” Sometimes a vehicle is “scared to cross railroads,” doesn’t know what to do at malfunctioning traffic lights or loops infinitely, such as when construction blocks the road.
Bryant Walker Smith, an autonomous vehicles researcher and associate professor of law at the University of South Carolina, said those are likely signs that the technology is not ready. As it stands, Tesla’s robotaxi service in Texas seems to be mostly a publicity move or investor relations showcase, he said.
“Tesla, which by all indications is a confident company run by a supremely confident person, is clearly not confident in the reliability of their technologies,” he said. “The scale of this is very small.”
Recent service expansions have been incremental. Last week, Tesla’s head of artificial intelligence, Ashok Elluswamy, said in a post on X that its driverless vehicles were now available to be hailed in the entire Austin geofence, after operating them in only a limited area since January. It was not immediately clear whether the change meant all robotaxis were monitorless. Tesla did not respond to a request for comment.
Musk has blamed state and federal regulators for the slow rollout. But Tesla has faced few rules in Texas, and it has moved slowly even in places where it has the green light. The company secured test permits in Arizona and Nevada last year but has yet to launch in those states.
Tesla’s halting progress is particularly evident in California. Although the company touts that it has robotaxis in the San Francisco Bay Area, they’re all operated by humans.
Behind the scenes, Casey Blaine, the automaker’s senior regulatory counsel, acknowledged to state regulators that the California service doesn’t qualify as driverless.
Its vehicles there use supervised driver-assistance software that is “not designed to operate without the presence and monitoring of a human driver,” unlike those in use in Texas, Blaine told the California Public Utilities Commission in a November email, which was obtained by Bloomberg News via a public records act request. Despite launching in California last July, Tesla still hasn’t taken steps that would allow it to offer an unsupervised driverless robotaxi service, representatives for the DMV and CPUC said.
That Tesla’s California rides are offered through the same robotaxi app used in Texas indicates the company’s driverless ride-hailing efforts are still mostly about optics, Smith said.
“There is a reality,” he said, “and Tesla is telling two diametrically opposed fictions about that reality.”
Tesla is also working to build out the unflashy parts that come with running a robotaxi business. Rideshare fleets require significant maintenance infrastructure to ensure the cars are clean and running properly.
On an April call with investors, Musk said the company is currently using existing charging stations and service centers for its robotaxis but will eventually need dedicated locations.
Tesla appears to be laying that groundwork now. In Austin, the company has filed plans with the city to expand an existing service center to include a “robocar” car wash and charging stations, with upgrades expected to cost more than $2 million.
A similar project appears to be taking shape in Las Vegas, with plans submitted for a Cybercab car wash. In the Phoenix metro area, the company has filed for at least two new charging stations for private company use, in Chandler and Mesa.
And in the Dallas-Fort Worth area, Tesla is working to obtain permits for a 35,000-square-foot service and charging center to be used as a robotaxi hub. The vehicles will get daily cleanings and service checks at that location, a Tesla representative said at an Irving City Council meeting last month.
The regional efforts represent crucial steps if Tesla wants to seriously operate a thriving commercial rideshare business. It’s likely the company will need similar setups as it looks to expand to cities that include Miami, Orlando and Tampa.
In the driverless ride-hailing race, Tesla has significant ground to make up, said Matthew Wansley, a professor at the Cardozo School of Law who specializes in autonomous technologies. He predicted the company will continue to rely on “crutches” for some time as it works to improve the service.
“I don’t want there to be a Waymo monopoly — I want there to be competition,” Wansley said. “But Tesla just hasn’t shown that.”
Carlson writes for Bloomberg.