SAN FRANCISCO — Uber and Pony.ai announced Friday an expansion of their autonomous vehicle partnership that will place more than 2,000 Pony.ai robotaxis across Europe, adding four cities to the one service already running in Croatia's capital Zagreb. The companies also extended the partnership to the Middle East, though they named no specific cities in either region and set no public timeline for the new deployments.
The Zagreb service, launched in late March, is the first commercial robotaxi operation in Europe according to the two companies. That single-city pilot is now the template for the broader rollout—a sequence that mirrors how Waymo built U.S. density by proving one market before expanding to the next.
Fleet size is the central economic variable in the robotaxi business. More vehicles on the road generate more trip data, which regulators require before approving wider commercial operations, and more data accelerates the training loops that reduce per-mile cost. Waymo currently operates roughly 5,000 vehicles, primarily in the United States, and holds the global fleet leadership position. Uber and Pony.ai's 2,000-vehicle European target would make it the largest autonomous fleet on the continent if deployed.
Waymo is not standing still on Europe. The Alphabet-backed company is testing rides in London and in June set up new legal entities in four major European Union economies—a structural move that typically precedes commercial licensing applications. Waymo has also started formal engagement with Japanese officials in Tokyo, where it has expansion plans underway.
Uber's Japan push is running through a different structure. On Thursday—one day before the Pony.ai announcement—Uber's Japanese subsidiary signed an agreement with a local taxi operator to handle daily fleet management for a robotaxi test in Tokyo later this year. That arrangement keeps Uber as the demand aggregator and booking layer while a local partner absorbs operational complexity, the same asset-light model Uber uses in most markets.
The competitive picture in Europe includes Chinese operators as well. Baidu's Apollo Go and WeRide are both expanding European tests. Uber already has a separate agreement with WeRide: in June, the two companies announced plans for Spain's first robotaxi pilot in Madrid, scheduled for later this year. Uber and WeRide also operate together in Abu Dhabi and Dubai, which gives Uber's Middle East extension with Pony.ai a market where it already has autonomous vehicle operational experience.
Uber CEO Dara Khosrowshahi said the company's goal is to become the leading commercialization platform for autonomous vehicles globally, describing a core piece of that strategy as assembling a broad dataset from across Uber's partner network that it can share back with those partners to advance vehicle development. The remarks came on an earnings call earlier this month, transcribed by FactSet.
Pony.ai is listed on the Nasdaq under the ticker PONY and is scheduled to report quarterly earnings Tuesday. The announcement of the European expansion came Friday, four trading days ahead of that release—timing that gives investors a concrete operational update to weigh alongside the financial results.
The economics of Uber's autonomous vehicle partnerships differ sharply from its human-driver model. Uber earns a platform fee on each ride without carrying driver costs or vehicle depreciation. As robotaxi unit economics improve with scale, the margin profile of those rides rises. The risk sits with the vehicle operator—Pony.ai in this case—which must fund hardware, maintenance and safety validation across multiple regulatory environments simultaneously.
Regulatory approval pace is the real constraint on the 2,000-vehicle figure. European transport regulators operate country by country, with no single EU-wide licensing framework for autonomous commercial vehicles. The Zagreb service in Croatia represents one approval in one national market. Each of the four unnamed cities will require its own regulatory process, meaning the timeline from announcement to operating fleet depends on factors neither company controls.
Pony.ai faces that regulatory burden with Waymo moving in parallel, Baidu and WeRide running their own European programs, and the clock ticking on first-mover advantages in specific city markets. The company that locks up the most favorable regulatory relationships in the highest-density European cities—not the one that announces the largest fleet number—ends up with the durable commercial position.
