Tesla’s Cybercab goes into production — so why is Musk tapping the brakes?

At a glance:

• Tesla’s Cybercab has entered continuous production at Gigafactory Austin, Texas, with initial builds dating to February and a public rollout video captioned “Purpose built for autonomy.”
• Expansion to Dallas and Houston is limited to two vehicles per week per city, with Elon Musk citing rigorous validation as the limiting factor and pledging no accidental injuries during scale-up.
• Musk tempered near-term expectations on the earnings call, acknowledging slow initial volumes through year-end and that millions of Hardware 3-equipped Tesla vehicles will need serious retrofits to reach unsupervised Full Self-Driving.

A guarded launch in Austin and beyond

Tesla’s Cybercab is now in production at the company’s Gigafactory in Austin, Texas, marking a milestone that caps years of ambitious promises about autonomy. The start of production was announced Thursday on X, accompanied by a video shot from inside a steering wheel-less Cybercab as it drove out of the factory with the caption “Purpose built for autonomy.” The company built a handful of initial Cybercabs back in February, but continuous production only began this month, underscoring the transition from prototype to low-rate manufacturing.

Despite the symbolic progress, the rollout remains tightly constrained. In recent remarks, Musk sounded uncharacteristically cautious about expansion, offering no new details about recent moves into Dallas and Houston. Each of those cities currently hosts only two vehicles per week after launch, a pace that highlights the gulf between Tesla’s earlier hyperbole and its present operational reality. Musk framed the limitation as a safety imperative rather than a logistical shortfall, insisting rigorous validation is the true bottleneck.

Validation vs velocity on the robotaxi frontier

“The limiting factor for expansion is really rigorous validation, making sure things are completely safe,” Musk said in response to questions about the slower-than-expected rollout. “We don’t want to have a single accidental injury with the expansion of Robotaxi, and we have, to the credit of the team, not had a single one to date.” That claim, however, sits uneasily alongside public data. Tesla has reported 14 crash incidents involving its robotaxis to the federal government since the service launched in Austin, Texas, a year ago. Unlike other robotaxi operators that disclose crash details and injury outcomes, Tesla routinely redacts that information, leaving outside observers unable to verify the safety profile the company asserts.

The tension between caution and candor is compounded by Tesla’s historical pattern of aggressive timelines. For years, Musk portrayed unsupervised Full Self-Driving as perpetually just around the corner, pointing to the success of Autopilot and then FSD (Supervised) as proof of an inevitable shift from human-driven to AI-piloted mobility. He has even invoked Tesla’s humanoid robot efforts to argue that autonomy is domain-agnostic and therefore unstoppable. Yet hundreds of crashes involving Tesla vehicles using FSD and Autopilot, along with dozens of deaths, have triggered multiple government investigations and placed a major FSD recall in the wings, creating a backdrop of regulatory and reputational risk.

Caps, controls and the compliance question

Another structural headwind for the Cybercab is its lack of conventional controls. Without a steering wheel, pedals, mirrors or other features required under Federal Motor Vehicle Safety Standards, Tesla must rely on exemptions that cap production at 2,500 vehicles per company. Legislation to lift that cap has languished in Congress for years, leaving purpose-built autonomous vehicles in a regulatory limbo that constrains scale. When asked on X whether the Cybercab’s production would be subject to the cap, vice president of Vehicle Engineering Lars Moravy replied “No,” indicating the company plans to self-certify compliance with existing standards, much as Amazon’s Zoox did with its purpose-built autonomous shuttles.

That path is not without precedent or peril. The National Highway Traffic Safety Administration opened an investigation into Zoox’s self-certification claim under the previous administration, though the probe was closed after the change in leadership. For Tesla, the strategy allows higher potential volumes but also invites scrutiny, particularly as the Cybercab is designed to operate without any human intervention. Until unsupervised FSD is demonstrably robust, the absence of fallback controls leaves little margin for error, a reality that may help explain Musk’s newly restrained tone.

Timelines, hardware limits and the S-curve reality

On the earnings call, Musk waffled between caution and overpromising, sketching an S-curve trajectory that has become a familiar rhetorical device. “Whenever you have a new product with a completely new supply chain, new everything, it’s always a stretched-out S curve, so you should expect that initial production of Cybercab and Semi will be very slow, but then ramping up, and going exponential towards the end of the year and certainly next year,” he said. He pledged to ramp up production of all vehicles in all factories to the best of Tesla’s ability through the balance of the year, but offered no firm guidance on when Cybercab volumes would meaningfully rise.

The software side is just as fraught. Musk said Version 15 of FSD, described as “a complete overhaul of the software architecture,” is coming by the end of this year or early next. Yet he also conceded that millions of Tesla vehicles equipped with Hardware 3 computers — sold between 2019 and 2023 — will not achieve unsupervised driving without serious retrofits, contradicting earlier commitments to future-proof capability. That admission narrows the plausible path to unsupervised service and raises questions about how many existing owners will pay to upgrade or wait for new vehicles instead.

What comes next for robotaxi reach and revenue

Last year, Musk claimed that by the end of 2025, 50 percent of the U.S. population would have access to Tesla’s Robotaxi service, calling the expansion “hyper exponential.” Today, the company operates in only a handful of cities, including Austin, Dallas, and Houston, while running a human-driven, invite-only ridehail service in San Francisco. Musk acknowledged on the call that “unsupervised FSD or Robotaxi revenue will not be super material this year,” but expressed confidence it will become “material probably in a significant way next year.”

Whether that confidence is warranted depends on more than factory throughput. It hinges on validating safety at scale, navigating the 2,500-unit exemption limit or securing regulatory relief, and delivering a step-change in FSD capability without further eroding trust. For now, the Cybercab’s production start is real, but the brakes remain half-on, a tacit admission that autonomy’s hardest mile is still ahead.