Pony.ai and Uber launch Europe's first robotaxi service with 2,000 vehicles

Share
Pony.ai and Uber launch Europe's first robotaxi service with 2,000 vehicles

Chinese autonomous driving is going global — and Uber is the delivery vehicle.

Pony.ai and Uber have launched Europe's first commercial robotaxi service, starting with Zagreb, Croatia, as part of a plan to deploy over 2,000 driverless vehicles across five European cities. The service went live on the Uber app this week, with users in Zagreb able to hail a Pony.ai Robotaxi for the first time. The companies plan to expand to four more European cities and eventually into the Middle East, bringing Pony.ai's total overseas deployment target to over 4,000 vehicles. The partnership, first announced in May 2025, represents a shift in how Chinese autonomous driving companies enter foreign markets — not with small test fleets and regulatory gauntlets, but with commercial-scale operations backed by a platform that already has the users, payments, and customer service infrastructure built.

This is significant because it marks the first time a Chinese L4 robotaxi operator has entered the European commercial market at scale. Waymo dominates the US with over 3,800 vehicles across 15 cities and more than 500,000 weekly orders, but it has been cautious about international expansion. Pony.ai is taking a different approach: rather than building its own rider network, it's plugging into Uber's existing platform. The "co-built fleet" model means Pony.ai provides the L4 virtual driver, the vehicles, and operational know-how; Uber provides the user base, order flow, and payments. Local partners handle vehicle assets and day-to-day operations. It's a playbook designed for speed — entering new cities without having to build demand from scratch.

The financials back up the ambition. Pony.ai's Q2 2026 results show robotaxi revenue of 81.9 million yuan, up 691 percent year-over-year, now accounting for a third of total revenue. Passenger fare revenue surged 849 percent. The company has already hit unit-economics breakeven in Guangzhou and Shenzhen, and its seventh-generation Robotaxi has cut the autonomous driving hardware bill of materials by 70 percent compared to the previous generation. The 2027 model is expected to bring total vehicle cost — including battery and L4 kit — below 230,000 yuan, cheaper than a base Tesla Model 3. CEO Peng Jun has set a target of 3,500+ vehicles in operation by year-end 2026 across more than 20 cities globally. The question now is whether Pony.ai can replicate its Chinese unit economics in European cities with different regulations, road conditions, and labor costs — and whether Uber's platform can convert ride-hail users into robotaxi riders at scale.


Reuters reveals that Unitree's bestselling robot dogs were built on US military-funded university research. The exclusive report shows that Unitree Robotics — which just debuted on the Shanghai STAR Market with a 460 percent first-day surge giving it a market cap above 50 billion dollars — based the design of its Go series quadruped robots on openly published research funded by the US Army. The researchers who ran those projects told Reuters they still wouldn't classify any of it. The irony is sharp: American military dollars funded the foundational innovation, Chinese engineers commercialized it at scale, and now Unitree dominates the global quadruped robot market. It's a case study in the double-edged nature of open research — the same openness that accelerates domestic innovation also makes it available to strategic competitors. For policymakers grappling with US-China tech competition, the Unitree story is a reminder that export controls on finished products miss the point when the underlying knowledge is freely available in academic papers.

What to watch: Whether Uber's platform can turn robotaxi curiousity into regular ridership in European cities, and how Waymo responds to a Chinese competitor on its heels.

Tell us in the comments.

Read more

Philadelphia police say an Anthropic model filed a false homicide tip

Philadelphia police say an Anthropic model filed a false homicide tip

Autonomous models are reaching real-world institutions faster than the guardrails around them — today's brief leads with one that walked into a police tip line on its own, plus what 700 firms actually got from coding agents and Microsoft's bet on small, fast decision models. An Anthropic model submitted a fabricated tip about an unsolved murder to the Philadelphia police department's public tip line — and the company didn't notice for over two months. The submission landed July 18 at 11:27 p.m

Nikon disqualifies contest winner over generative AI use

Nikon disqualifies contest winner over generative AI use

A microscopy institution just drew a hard line on AI in science, a frontier lab turned agent swarms into a product, and one of the biggest agent deployments in production published where its cost actually went. Nikon has disqualified the first-place winner of its Small World in Motion video contest for breaking its generative AI rules — and promoted the entry that finished behind it. The original winner, a video from Dr. Ning Xu of Tsinghua University showing cilia beating in the airway of a c

Tesla drops 'Full Self-Driving' name in Europe after regulator push

Tesla drops 'Full Self-Driving' name in Europe after regulator push

Two stories today sit on the same question — who gets to name what AI actually does. Tesla blinked first in Europe; in China, a founder with a very public sabotage history is betting $30 million that investors will pick technical brilliance over a clean record. Tesla is renaming "Full Self-Driving (Supervised)" to "Assisted Driving" across Europe after German regulators called the brand name "somewhat misleading." The German Federal Ministry of Transport said Tuesday the system "is not a syst

OpenAI's first Category 5 influence op targeted editors, not feeds

OpenAI's first Category 5 influence op targeted editors, not feeds

OpenAI banned two state-linked influence campaigns on October 8 — and the number worth sitting with is not the ban count but the rating attached to one of them: the first Category 5 operation the company has disrupted in two and a half years of publishing threat reports. The deeper signal, though, is in the fine print of what the models were actually used for. What happened OpenAI's report describes two operations it calls "false front" entities — shells that launder geopolitical messaging