Pony.ai and Uber plan more than 2,000 robotaxis across four European cities

Why it matters

Pony.ai is using Uber's distribution and local fleet capital to pursue European scale without carrying every operating cost, while permits, launch dates and Uber-app availability in the four new cities remain unresolved.

Pony.ai co-founders James Peng and Tiancheng Lou plan to put more than 2,000 robotaxis into four additional European cities through an expanded agreement with Uber, with local operators handling fleet work and partners potentially owning the vehicles.

The expansion was announced by Uber and Pony.ai on August 14, 2026 and reported by TechCrunch that day. It builds on Pony.ai's commercial service in Zagreb, Croatia, which began in April.

The 2,000-vehicle target is a plan without a deployment calendar, named cities or a city-by-city vehicle allocation. Uber and Pony.ai also left the supervision model undefined, including how many vehicles would operate fully driverless and how many might begin service with human safety personnel.

Those gaps matter because Peng is pursuing a large expansion through what Pony.ai calls a joint-deployment model. Pony.ai provides the Level 4 autonomous-driving system and operating expertise, while Uber integrates the service into its ride-hailing network. Local providers can manage maintenance, cleaning and charging. The companies said different partners may own the robotaxi fleets depending on the market, without identifying Uber as a fleet owner.

The structure could let Peng put Pony.ai's software into more cities without requiring Pony.ai to operate every vehicle itself. It also gives Uber another autonomy supplier for a marketplace designed to mix human drivers and robotaxis.

Peng and Lou built Pony.ai around the driving stack

Peng spent six years as chief architect for Baidu's autonomous-driving unit after working on advertising systems at Google, where Pony.ai says he received the Google Founder's Award. He holds degrees from Tsinghua University and the State University of New York at Buffalo, as well as a doctorate from Stanford University, according to Pony.ai's management biography and securities filings.

Lou, Pony.ai's chief technology officer, studied computer science at Tsinghua before joining Google in 2012. He worked in Google's autonomous-driving group before holding roles at Quora and Baidu, where Pony.ai's annual filing identifies him as Baidu's youngest principal engineer.

Peng and Lou founded Pony.ai in Fremont, California, in late 2016. Pony.ai now describes its core product as the "Virtual Driver," a vehicle-independent package of software, computing hardware and related services. The European agreement tests whether that technology can become a repeatable product for fleet owners operating under different local rules, rather than a service Pony.ai must build city by city on its own balance sheet.

The operating model behind the expansion is straightforward: Pony.ai supplies the technical layer, while partners provide distribution, vehicles and local execution.

Zagreb shows the work left between announcement and scale

Zagreb offers an early version of the arrangement. Uber, Pony.ai and Croatian autonomous-mobility company Verne announced the project on March 26, assigning Pony.ai the driving system, Verne fleet ownership and local service operations, and Uber integration with its ride-hailing network. Peng said the partnership combined Pony.ai's autonomous-driving technology, Uber's platform and Verne's deployment role to build a scalable and sustainable autonomous-mobility operation.

Pony.ai said commercial service began on April 8, operating from 7 a.m. to 9 p.m. across roughly 90 square kilometers, including Zagreb Airport. Riders initially booked and paid through Verne's app. The April announcement said availability through Uber's app was coming later.

As of August 14, Pony.ai and Uber were operating the Zagreb service and had announced expansion to four additional European cities. The August announcement did not state whether booking through Uber's app had gone live. Zagreb has established that Pony.ai vehicles can carry paying passengers in Europe through a local fleet owner and service operator, though the public announcements have yet to document the Uber-app distribution model planned for the larger rollout.

Pony.ai has targeted more than 3,000 robotaxis across over 20 cities by the end of 2026. A European plan exceeding 2,000 vehicles would be large relative to Pony.ai's reported fleet, although the lack of a deployment date means the two figures cannot be treated as part of the same timetable.

Local vehicle ownership is what makes the target conceivable. It also makes the announcement harder to measure. A partner-owned vehicle carrying Pony.ai's autonomy system counts toward deployment without requiring Pony.ai to purchase the car. Fleet ownership, operating responsibility, permits and driverless approval can vary by city, leaving the 2,000 figure dependent on several parties beyond Peng and Lou's control.

Uber wants suppliers, Pony.ai wants distribution

Uber has been assembling a portfolio of autonomous-driving partners instead of tying its network to a single technical system. In February, Uber introduced Uber Autonomous Solutions, packaging services such as demand generation, fleet support and insurance for autonomy developers.

Uber has since announced separate deployment plans with Rivian and Nvidia, among others. Each agreement increases Uber's access to robotaxi supply while making autonomy developers compete for riders and fleet partners inside Uber's marketplace.

For Pony.ai, Uber offers access to an established ride-hailing network without forcing Peng and Lou to recreate that distribution in every country. Pony.ai becomes one supplier among many, with less control over the customer relationship and no guarantee that Uber will favor its vehicles when competing autonomous systems are available.

European regulation remains the constraint that distribution and fleet financing cannot remove. Pony.ai's 2025 annual report warns that permits, commercialization delays, export controls and international trade restrictions can limit where Pony.ai operates or obtains technology. Each new city will require local regulatory work of the kind Verne handled in Zagreb.

The agreement gives Peng a capital-conscious route toward the international scale Pony.ai has promised investors. Execution will be visible in permits, vehicles carrying paying passengers and documented availability through Uber's app. Until those arrive in the four additional cities, the 2,000 robotaxis remain a target divided among unnamed markets.