Illustration: AI-generated · Inside China AI
The Robotaxi Math Is Solved — and It’s Arriving in Europe
For years, the robotaxi story was a money-burning machine with good PR. That era is ending — and the proof is arriving in Europe. On August 14, Pony.ai and Uber announced an expanded partnership to deploy more than 2,000 robotaxis across Europe, growing from their commercial service in Zagreb to four additional European cities. The reason this announcement deserves more attention than the usual pilot-project news: Pony.ai is scaling into Europe after proving its robotaxis can make money, not in the hope that they someday will.
The math behind the move
Pony.ai operates fully driverless, paid robotaxi services in four Chinese cities — and the unit economics have flipped. The company reached per-vehicle profitability in Guangzhou in November 2025 and in Shenzhen in February 2026. Robotaxi revenue grew roughly 400 per cent year-over-year in Q1 2026. And the cost curve keeps falling: Pony.ai plans to push the manufacturing cost of its seventh-generation robotaxi platform below 230,000 yuan — roughly $33,000 — by 2027, about the price of a mid-range electric car. The fleet is set to exceed 3,000 vehicles by the end of 2026, across more than 20 cities worldwide.
Why Zagreb, why Uber
The European playbook is deliberate. Pony.ai established its first European hub in Luxembourg, where it received an autonomous-vehicle testing permit in April 2025, and launched commercial service in Zagreb this spring with Croatian mobility operator Verne. Croatia offered what larger markets could not: a receptive regulator and a real city to prove the model. Uber brings the missing piece — a platform with over 200 million users, so Pony.ai never has to build European consumer demand from scratch. “This expanded agreement marks an important new phase in the partnership between Pony.ai and Uber,” said founder and CEO James Peng. The four expansion cities remain unnamed; Uber has previously pointed to London, Madrid, and Munich as future robotaxi markets.
Pony.ai is not alone — this is a wave
Zoom out, and August 2026 looks like the month Chinese autonomous driving arrived in Europe in force. WeRide announced a Danish deployment in early August, extending a European footprint that spans six countries. Baidu’s Apollo Go began testing in London in late July. Momenta holds a Level-4 approval in Germany. Uber itself hedges across the field, partnering with Baidu, WeRide, Wayve, and others. The pattern mirrors what we documented in humanoid robotics: Chinese companies iterate at home at enormous scale, drive costs down, then arrive in Europe with proven economics while local alternatives are still in the pilot phase. The robotaxi wave, company by company: our profiles of WeRide and Apollo Go.
Why it matters for Europe
Europe has no domestic robotaxi champion at scale — and its streets are becoming the export market where Chinese autonomy proves itself against Waymo’s America. That creates a genuine dilemma for European policymakers: welcome the technology and cede the operational layer of urban mobility to Chinese-American platform pairings, or slow it down and explain to citizens why robotaxis work in Zagreb, Shenzhen, and Atlanta but not in Berlin. The Zagreb model shows how the decision is actually being made: smaller EU states are moving first, creating regulatory facts while the large member states deliberate. For European mobility operators, the question is no longer whether to engage with Chinese AV technology — it is whether to be a Verne, operating the fleet, or a bystander.
No responses yet