Tesla Robotaxi vs Waymo: One Has the Hype, the Other Has the Miles
Who is winning, Tesla or Waymo? As of mid-2026, Waymo leads on every metric that matters: about 500,000 paid driverless rides a week across 11 US cities, more than 200 million rider-only miles, and a published safety record of 92% fewer serious-injury crashes than human drivers. Tesla's robotaxi is real but tiny, roughly 40 cars in 3 Texas cities, and its bet is different: dirt-cheap, camera-only hardware that could scale to millions of cars later. Today Waymo is ahead. The open question is whether Tesla's cost advantage closes the gap.
Two companies are running cars with nobody behind the wheel and charging you for the ride. That is where the similarity ends. Waymo, the quiet Alphabet subsidiary that has been at this since the Obama administration, now does roughly half a million paid driverless trips every week. Tesla, which launched its long-promised robotaxi in Austin in June 2025, is running about forty cars. Both are called robotaxis. Only one of them is a business yet.
That gap is the real story of the driverless race in 2026, and it is more interesting than the usual "Elon versus everyone" framing. Because the two companies did not just build different cars. They made opposite bets on what a self-driving car even is.
Two bets on the same future
Waymo's car is a sensor fortress. The sixth-generation Waymo Driver carries 13 cameras, 4 lidar units, and 6 radars, all feeding a system that cross-checks every object three different ways and leans on high-definition maps it has built street by street. It is expensive, it is heavy, and it works in the rain.
Tesla threw all of that out. The Cybercab and the Model Ys running in Austin use cameras only. No lidar, no radar, no pre-built maps. Tesla's argument is that humans drive with two eyes and a brain, so a neural network with eight cameras should manage too, and that the moment it does, every Tesla already on the road becomes a potential robotaxi overnight. The hardware on a Cybercab costs a rounding error next to Waymo's sensor stack. That is the whole point.
One bet is "make it work perfectly, then make it cheaper." The other is "make it cheap, then make it work." In 2026 we are finally watching both collide with reality.
Where they actually operate today
This is where the hype meets the map. Waymo runs paid, fully driverless rides in 11 US cities, including San Francisco, Phoenix, Los Angeles, Austin, Atlanta, Miami, and a wave of Texas metros it opened in early 2026. Its fleet is around 3,800 vehicles and climbing, with a new Chinese-built Zeekr van and a 50,000-unit Hyundai Ioniq 5 deal coming online.
Tesla operates in three: Austin, Dallas, and Houston. Texas records show about 42 driverless vehicles authorized as of late May 2026, and the fleet has barely grown while Tesla waits on its next-generation FSD v15 software. For context, Elon Musk had promised hundreds of cars and coverage for half the US population by the end of 2025. The real number a few months later was a few dozen cars in Texas. The gap between the promise and the parking lot is the most honest measure of where Tesla actually is.
Waymo runs away with scale, on cities and on fleet size. Tesla's one edge is cost: a Cybercab is meant to cost roughly half what a sensor-laden Waymo does. Tap "Cost per car" and the winner flips. That single number is the entire Tesla bet.
The safety record nobody can spin
Numbers like rides-per-week can be waved away as a head start. Safety cannot. Waymo has published peer-reviewed data covering more than 170 million rider-only miles: 92% fewer serious-injury-or-worse crashes than human drivers on the same roads, 82% fewer injury crashes, and 92% fewer crashes that hurt pedestrians. That is the single most important number in this entire debate, and it belongs to Waymo.
Tesla's early record is rougher. Independent tallies of federal crash reports found the robotaxi involved in crashes at a rate several times higher than an average human driver over its first eight months, and the service still pauses in heavy rain because camera-only vision degrades when the lenses get wet. Tesla initially redacted the crash narratives, which were only made public under pressure in 2026. None of this means camera-only driving cannot work. It means it does not work as well yet, and Tesla is running it on public streets while it improves.
Waymo vs Tesla at a glance
Strip out the noise and put the two side by side. One of these columns is a scaled-up business; the other is an experiment with enormous potential.
| Measure | Waymo | Tesla Robotaxi |
|---|---|---|
| First paid service | October 2020 (Phoenix) | June 2025 (Austin) |
| Cities live | 11 US cities | 3 (Austin, Dallas, Houston) |
| Fleet size | ~3,871 vehicles | ~42 driverless |
| Vehicles | Jaguar I-Pace, new Zeekr, Ioniq 5 | Model Y (Cybercab in production) |
| Sensors | Lidar + radar + cameras | Cameras only |
| Supervision | Fully driverless | Driverless in Austin |
| Est. cost per car | ~$60,000 | ~$30,000 (target) |
| Safety record | 92% fewer serious-injury crashes than humans | Early rate above human average |
The Cybercab wildcard
Here is why Tesla is still in this. The purpose-built Cybercab entered production at Giga Texas in 2026: a two-seat pod with no steering wheel and no pedals, an EPA-rated efficiency around 5.5 miles per kWh, and a target price near $30,000. A Waymo costs an estimated $60,000-plus per car even after the latest sensor diet. If Tesla's software ever crosses the reliability line, it can flood the market with robotaxis at half the unit cost, and let ordinary owners add their cars to the fleet. Waymo has the better car today. Tesla has the cheaper one, and a far bigger potential army of them.
The catch is that "if the software crosses the line" is carrying the entire argument, and it has been carrying it for years. The Cybercab is in production but was not yet carrying paying passengers as of mid-2026. It is legal to build. It is not yet approved to drive itself for hire anywhere.
How the robotaxi race got here
The two companies arrived at the same June 2026 moment from completely different directions and speeds.
Tesla unveils the Cybercab at its "We, Robot" event and promises a roughly $30,000 price.
Waymo opens paid rides in Austin, its fifth city, while still expanding on the West Coast.
Tesla launches its robotaxi in Austin, with safety monitors riding along at first.
Tesla removes the safety monitor in parts of Austin: its first genuinely driverless rides.
Waymo passes 200 million driverless miles, starts its sixth-gen rollout, and raises at a $126B valuation.
Tesla opens Dallas and Houston and mass-produces the Cybercab; Waymo's first purpose-built Zeekr robotaxi takes paying riders.
So who is winning?
If winning means a real, paid, driverless business operating at scale with a safety record to defend, it is not close: Waymo is winning, and it spent fifteen years and a fortune earning that lead. If winning means owning the cheapest path to a million robotaxis, Tesla still holds the more dangerous long-term weapon, assuming the camera-only gamble pays off.
It helps to be precise about what each company has actually proven. Waymo has shown that a car can drive itself safely, in real cities, for paying strangers, millions of times. Tesla has shown that it can build an autonomous-looking car very cheaply and get a small fleet onto public roads fast. Those are different achievements, and only one of them is a business today. The way to tell who is really winning over the next year is not the next demo or the next tweet. It is two boring questions: does the Cybercab start carrying paying riders, and does Tesla's crash rate fall as the software matures?
The EV-Global Verdict
Waymo wins the present on miles, cities, and safety. Tesla owns the cheapest possible future, but only if FSD finally delivers what Musk has promised since 2019. The smart way to watch the rest of 2026 is to ignore the noise and track two things: whether the Cybercab actually carries paying riders, and whether Tesla's crash rate drops as FSD v15 ships. Until both of those turn green, if a car pulls up with no driver and you want the safe bet, you want the one with the spinning lidar on the roof.
Tesla Robotaxi vs Waymo: frequently asked questions
Is Tesla's robotaxi fully driverless?
Yes, in parts of Austin since January 2026, with no safety monitor in the car, though it still pauses in heavy rain. Dallas and Houston launched driverless in April 2026.
How is Waymo so far ahead of Tesla?
A roughly 15-year head start, lidar-plus-camera redundancy, and high-definition maps. Waymo reached paid driverless rides years before Tesla and now runs about 500,000 a week across 11 cities.
What is the Tesla Cybercab?
A purpose-built two-seat robotaxi with no steering wheel or pedals, in production at Giga Texas since 2026, targeting roughly 30,000 dollars, which is far cheaper than a Waymo vehicle.
Is Waymo or Tesla safer?
On published data, Waymo reports 92% fewer serious-injury crashes than human drivers over more than 170 million rider-only miles. Tesla's early robotaxi crash rate ran higher than human drivers.
Why does Tesla use cameras only?
To slash hardware cost and let existing Teslas join the fleet later. The trade-off is weaker performance in rain, fog, and low-sun glare, which is why the service pauses in bad weather.
Can I buy a Tesla Cybercab?
Tesla has shown a roughly 30,000 dollar target and started production, but the Cybercab is built for fleet and robotaxi use. Wide consumer availability and approval to drive itself for hire are still pending.
Photo: Maxim / Unsplash. Resized and converted to AVIF.
