Three years ago, robotaxis were a niche experiment in Phoenix and Wuhan. Today, Waymo completes around 500,000 paid rides per week, Apollo Go has passed 22 million cumulative rides, and Europe's first commercial robotaxi is driving in Zagreb.
Time for a sober look at the numbers. For this article, I compiled the official ride and safety data, the providers' quarterly reports, US regulatory documents, and Europe's plans, and had every key figure cross-checked. Including the gap between Tesla's announcements and Tesla's actual fleet.
- Waymo dominates the West: around 500,000 paid rides per week (March 2026), over 20M rides in total (officially confirmed as of late 2025), ~3,500 vehicles across 11 public cities, and a $16B funding round in February 2026
- Baidu's Apollo Go is the quiet giant: over 22M cumulative rides (April 2026) across 27 cities worldwide, now including driverless service in Dubai and Abu Dhabi
- Tesla's robotaxi reality remains small: Bloomberg counted 59 vehicles in June 2026, against a promise of 500+ for Austin alone. And Europe is only just getting started, with Zagreb as the premiere
1. The Market at a Glance
The most important metrics of the major providers:
Two companies play in a league of their own. Waymo leads on weekly rides and fleet size, Apollo Go on cumulative rides. Everyone else, Tesla included, trails by at least an order of magnitude.
2. Waymo: The Market Leader
2.1. Half a Million Rides per Week
Waymo's growth curve is the most impressive time series in the industry:
From 50,000 to 500,000 weekly rides in just under two years, a tenfold increase. In 2025, Waymo completed around 15M rides for the full year, three times as many as in 2024. Co-CEO Tekedra Mawakana has set a target of 1M rides per week by the end of 2026.
2.2. From Phoenix to Las Vegas: The Cities
Eleven cities are publicly bookable, and four more (Las Vegas, Denver, San Diego, Tampa) entered driverless employee-only operation in July 2026. International plans are next: London is set to become Waymo's first market outside the US in 2026, and Tokyo is in testing.
2.3. The $16 Billion Round
In February 2026, Waymo raised $16 billion, led by Sequoia, DST Global, and Dragoneer, at a $126 billion valuation. For comparison, its valuation stood at $45 billion in late 2024. The robotaxi market now has its first company in blue-chip territory.
3. Safety: What the Data Actually Says
The industry's most important selling point is safety. Waymo's Safety Hub compares its own crash record with a locally adjusted human-driver benchmark, now covering 220.6 million driverless miles (as of March 2026):
94% fewer crashes involving serious or fatal injuries, 93% fewer pedestrian injury crashes. Those are Waymo's own figures, but independent sources back them up:
A joint study with reinsurer Swiss Re (baseline: over 500,000 insurance claims) found 88% fewer property damage claims and 92% fewer bodily injury claims. And a manufacturer-independent University of Central Florida study in Nature Communications confirms lower crash likelihood for autonomous vehicles in most scenarios, with two exceptions: dawn/dusk lighting and turning maneuvers.
Even the market leader isn't flawless, though. More on that in section 8.
4. Tesla: Big Promises, Small Fleet
No robotaxi program generates more headlines than Tesla's, and none has a wider gap between announcement and reality.
The facts:
Tesla launched in Austin on June 22, 2025, with about 10 Model Ys and safety monitors in the passenger seat. In July 2025, Elon Musk promised to cover "half the US population" by the end of 2025; in the fall, a target of over 500 vehicles for Austin alone followed. In reality, Bloomberg counted 59 vehicles across three Texas cities in June 2026.
To be fair, there has been real progress:
Since June 2026, a growing share of the Austin fleet has been driving without safety monitors across the entire metro area (though in a Bloomberg field test, roughly one in three rides still had a monitor on board), and the Miami launch on July 3, 2026, was driverless from day one, Tesla's first market outside Texas and California. The Bay Area remains the curiosity. There, according to California's CPUC, Tesla's "robotaxi" legally operates as a limo service, with a licensed human driver as the operator. As of the latest reports, Tesla had not applied for an actual driverless permit.
5. Apollo Go: The Quiet Giant From China
While the West watches Waymo and Tesla, Baidu's Apollo Go has quietly built the world's largest cumulative ride count:
Over 22M rides (as of April 2026), and the pace is accelerating: Q4 2025 alone added 3.4M rides (peak weeks above 300,000), and by March 2026 the weekly record already topped 350,000. Apollo Go operates in 27 cities worldwide and is expanding aggressively abroad, driverless in Abu Dhabi since January 2026, in Dubai since April 2026, in the greater Seoul area since February 2026, plus trials in Switzerland (as "AmiGo" with PostBus). It doesn't always run smoothly, though: in late March 2026, an operational glitch in Wuhan left nearly 100 robotaxis stranded in traffic at once.
The most important number might be an economic one, though:
Baidu put the production cost of its sixth-generation vehicle (RT6) at around $28,600 when it unveiled the car in 2024, a fraction of Western robotaxi costs. And in Wuhan, Apollo Go has been profitable on a per-vehicle basis since late 2025, according to Baidu. Robotaxis will turn profitable in China first.
6. The Challengers: Zoox, WeRide, Pony.ai & Co.
The pace of the Chinese providers stands out. Pony.ai grew its fleet sixfold within a year, from under 300 to over 1,700 vehicles, and increased revenue by 395% in Q1 2026. WeRide holds licenses in eight countries. And Amazon's Zoox has already carried over 500,000 riders (June 2026) but can't charge money yet, pending a US regulatory exemption.
Fleet sizes side by side:
Apollo Go is deliberately missing from the fleet comparison: Baidu doesn't publish a total fleet figure, only per-city values from industry press, such as over 1,000 vehicles in Wuhan, its largest single market.
7. Prices: The Robotaxi Is Catching Up
Do robotaxis cost more than Uber? Yes, still. But the curves are converging (all values in USD):
Waymo got about 4% cheaper within nine months, while Uber rose 12% and Lyft 7%. Waymo's premium over Uber shrank to just over $2 per ride. Factor in the tip you don't pay, and the gap is effectively even smaller. Long term, the economics favor the robotaxi: no driver, falling hardware costs, rising utilization.
8. Regulation & Setbacks: The Year of Investigations
2026 is also the year regulators got serious. The most important open proceedings:
The contrast with California's test data is interesting. Waymo most recently drove 5.2M miles there with just one intervention (disengagement) every 63,415 miles, the industry's best value. The regulatory proceedings still show that even statistically safe systems produce individual incidents that regulators and the public won't let slide. The California DMV even adopted new reporting rules in April 2026 that replace the disengagement statistic, long considered outdated, with new safety metrics.
9. Market Forecasts: A Billion-Dollar Market With a Wide Range
Goldman Sachs offers the most detailed model: from around 7,000 robotaxis worldwide (2025) to roughly 1M (2030) and 6M (2035), with a global market volume of $415 billion. The spread across forecasts ($105 to $415 billion for 2035) comes, as usual, from different market definitions. The common denominator is clear nonetheless. Everyone expects fleet growth by a factor of 100+ within ten years.
10. Europe & Germany: The Late Start
In 2026, Europe finally got its first commercial robotaxi service, just not in Germany, but in Croatia.
Since April 8, 2026, Verne, Rimac's robotaxi subsidiary, has been driving in Zagreb with Pony.ai technology and Uber as a partner, across 90 square kilometers including the airport. Ambitious plans follow across the continent: Waymo wants to serve London in 2026, Uber/Wayve are already collecting waitlists there, WeRide and Uber have announced Madrid (launching toward the end of 2026, initially with safety drivers) and are planning Zurich, and Apollo Go is testing as "AmiGo" with PostBus in eastern Switzerland.
And Germany? The status here is soberingly concrete:
The legal framework has existed since 2021 (autonomous driving act) and 2022 (AFGBV), among the first countries worldwide. The biggest project is ALIKE in Hamburg: MOIA (VW) and HOCHBAHN are testing up to 20 autonomous ID.Buzz vans across 37 square kilometers, with €26M in federal funding, targeting regular operation in 2026 (still with safety drivers; a delay can't be ruled out). For Munich, Uber, Autobrains, and Nvidia announced a pilot program in June 2026, so far without a launch date. And be careful with Mercedes headlines: the "MB.Drive Assist Pro" system announced for late 2026 is a level 2 assistant that requires driver takeover, not a robotaxi.
11. Milestones: From 2009 to Today
What stands out is the long runway this industry needed. Eleven years passed between Google's project start in 2009 and the first public driverless service, and another five and a half between Phoenix and half a million weekly rides. The 2018 Uber fatality and the 2024 Cruise shutdown are reminders of how quickly a single incident can cost years of progress.
12. Conclusion: The Tipping Point Is Here, but Only in Places
2026 is the year robotaxis go from experiment to business. Waymo grew its rides tenfold in a year and a half, Apollo Go is approaching break-even in Wuhan, prices are converging with Uber, and the safety data is consistently positive across multiple independent sources.
At the same time, the market remains extremely concentrated. Two companies account for the bulk of all rides worldwide, Tesla remains a niche operator with 59 vehicles (as of June 2026) despite all the headlines, and Europe is only just beginning. Whether the tipping point becomes a mass market will be decided on two fronts: with regulators, who became noticeably stricter in 2026, and on per-vehicle costs, where China sets the pace.
For the AI models behind these vehicles, see my AI statistics. And for how autonomous systems are scaling beyond the road, check out my AI agent statistics.
