Tesla's robotaxi fleet hits 970: how the Cybercab stacks up against Waymo, what Munro found inside the factory, and what it means for Uber

Fifty-four days after Tesla started taking paying passengers in its steering-wheel-free Cybercab, the robotaxi is no longer a concept car. Tesla's Texas robotaxi fleet, which mixes Cybercabs with driverless Model Ys, reached 970 registered vehicles on 26 October. And a new factory walk-through with engineering teardown specialist Munro & Associates says the Cybercab line can already build cars faster than Tesla can put them on the road.

So how quickly is Tesla really growing, how does that compare with Waymo, and where does it leave Uber and the ride-share drivers who use it? We go through it below.

The 1,000 figure: what it does and doesn't mean

The "1,000 Cybercabs" headline needs some care, because three different numbers are in circulation:

• About 1,000 (estimated): the number of Cybercabs covered by the US safety regulator's audit. When NHTSA opened audit query AQ26002 on 3 September, it put the affected population at "1,000 (Estimated)". That's NHTSA's estimate of the cars in scope, not a count of cars on the road.
• 970 (reported): all Tesla robotaxis registered in Texas as of 26 October, 463 Cybercabs plus 507 Model Ys, from the Texas DMV's automated-vehicle roster. That's up from 546 at the start of October. This is Texas only. Tesla doesn't publish a California robotaxi count, and its California service still has a safety driver.
• 463 (reported): Cybercabs in the Texas robotaxi fleet as of 25 October, again from state registration records. That works out to about 8.2 new Cybercabs a day since launch. Tesla's own last official count was "300+ Cybercabs serving Austin" on 10 October, an eightfold increase in 30 days.

Registered isn't the same as carrying passengers at any one moment, and community trackers that count cars actually seen in service have been well below the registration totals. The honest summary: Tesla has 970 robotaxis registered in Texas, 48% of them Cybercabs, and it's adding them every day.

Waymo's timeline: years, not weeks

Waymo, Alphabet's self-driving company, is the benchmark. It took a very different path to scale:

Put simply, Waymo took about five and a half years to go from its first fully driverless public service to a fleet of around 3,000. Tesla's Cybercab went from zero to several hundred in under two months. But Waymo's cars run in 15 cities, carry around half a million paid trips a week, and have a long public safety record. Tesla's Cybercab runs in one geofenced city, and Tesla doesn't publish ride numbers.

The gap is narrowing for one reason: manufacturing. Waymo fits its hardware to vehicles other companies build: first Jaguar I-Pace SUVs, and now the Zeekr-built "Ojai" minivan, which helped its Texas fleet grow 49% in three weeks in September (reported). Tesla designs and builds the whole car itself, in a factory it owns.

Inside the Cybercab line: what the Munro tour showed

On 26 October, Munro Live published a long engineering walk-through of the Cybercab, led by Tesla vice president of vehicle engineering Lars Moravy, Eric Earley and other Tesla engineers. The parts that matter most for scaling up:

• The line is ahead of demand. Moravy: "We actually don't run the line every day because the line can run so fast it's currently outpacing our robotaxi deployments." Tesla's Q3 update lists installed Cybercab capacity at more than 125,000 a year.
• "Unboxed" assembly is real. The car is built as separate modules (front, rear, both sides, and a structural battery-and-seats unit), each checked on its own before they're joined in an automated framing cell. Wheels go on at the second-last station, so a Cybercab first touches the ground as it rolls off the line.
• Automation (reported from the video): about 78% overall, 100% in framing, 25 final-assembly stations, and roughly half the factory floor space per car. There's no conventional paint shop.
• Battery and running gear (reported from the video): 4680-cell modules in a steel enclosure that can be swapped for servicing, front-wheel drive tuned for efficiency, brake-by-wire with no hydraulic master cylinder, and a 500,000-mile (about 805,000 km) design life.
• Charging: a conventional charge port for now, so the cars can use the existing Supercharger network. Tesla said Optimus robots might plug them in eventually. Inductive charging isn't in the near-term plan.
• Doors: asked why it chose butterfly doors that could cost more to repair, the Tesla team said, first of all, that they look cool. They also said the doors make it easier to get in and out.

We haven't independently checked the automation and floor-space percentages against a Tesla document. They come from the video as reported by people who watched it, so treat them as Tesla's claims. Costs: the publicly reported target is a consumer price of about US$30,000 (about A$46,000, est.) and a running cost of about 30 US cents a mile. Tesla hasn't confirmed a production cost.

The brake on growth: regulators

If the factory isn't the limit, approvals are. NHTSA's audit asks how Tesla certified a car with no steering wheel, pedals or mirrors as meeting federal safety standards, and in particular which standards Tesla decided didn't apply. Tesla's sworn answers are due on 30 October. Each new city also has its own rules, which is why Dallas, named for November, is only the second Cybercab market. At the Q3 results, Tesla said it wants Cybercabs in customers' hands by the end of 2026, though Elon Musk said most will be Tesla-owned and run on its own ride-hailing network.

What it means for Uber, Lyft and drivers

Uber's response is to become the marketplace for everyone else's robotaxis. At its Q2 results in August it committed more than US$10 billion (about A$15 billion, est.) to autonomous vehicles over the coming years, and it says AV partners have pledged 120,000 robotaxis to its platform (reported). The biggest single bet is Lucid Gravity SUVs running Nuro's self-driving system. Uber has ordered at least 35,000 of them (reported), it's due to launch in the San Francisco Bay Area in late 2026, and Houston follows in mid-2027.

The risk for Uber is that the strongest players don't need it. Waymo has wound down its Phoenix arrangement with Uber and plans to run its own app in Austin and Atlanta from 2028 (reported). Tesla is running its own Robotaxi app. A company that makes and owns its cars, and has its own app, keeps the whole fare.

For human drivers, the change will come city by city, starting where the weather is mild and the regulators are friendly. Five years ago, Waymo was the only company running fully driverless public rides. Now several companies have commercial services in the US, and Tesla's factory claims suggest the vehicles themselves will soon stop being the bottleneck.

And in Australia?

No robotaxi service with nobody behind the wheel operates in Australia, and none of Tesla, Waymo or Uber has announced a date for one. Federal automated-vehicle safety laws are still being developed. Australians already pay for Tesla's FSD (Supervised), at $149 a month, but it legally needs an attentive driver. Uber, the biggest ride-share platform here, hasn't named Australia in any of its AV launch plans. For Australian ride-share drivers, the US numbers above are worth watching. They're not an immediate threat.