Uber and Pony.ai plan over 2,000 robotaxis across Europe, expanding from Zagreb

Partnership expansion

Uber and Pony.ai announced on Friday an expanded partnership to deploy more than 2,000 autonomous robotaxis across Europe, building on a commercial service already running in Zagreb, Croatia. The two companies, which first partnered in May 2025 with a focus on the Middle East, launched Europe's first fare-charging robotaxi service in Zagreb in late March 2026 together with Croatian mobility startup Verne. Rides in Zagreb are currently booked through the Verne app, with a safety driver present at launch who could intervene; the service will soon become available through the Uber app. The vehicles used are the Arcfox Alpha T5, a five-door electric compact SUV manufactured by Chinese automaker BAIC.

The expansion will bring robotaxis to four additional European cities, though the companies have not named them or provided a deployment timeline. Details will be revealed in phases, according to the announcement. The partnership also extends to the Middle East, where Uber and Pony.ai already collaborate in Abu Dhabi and Dubai.

Division of roles

Pony.ai, founded in 2016 and listed on the New York and Hong Kong exchanges, supplies Level 4 autonomous driving technology, which allows vehicles to operate entirely without a driver within designated areas and conditions. The Guangzhou-based company has produced roughly 1,700 robotaxis to date, working with manufacturers including Toyota, BAIC, and GAC rather than building vehicles itself. Uber provides its ride-hailing platform for booking, payment, and customer support, alongside its network of human drivers. Daily fleet operations, including maintenance, cleaning, and charging, will be handled by local partners in each market, and vehicle ownership may vary by location.

Executive statements

Pony.ai founder and CEO James Peng described the expansion as a step toward broader commercial operations.

This is a new important phase in the partnership, aimed at building large-scale sustainable commercial operations in Europe and beyond.

Uber's head of autonomous mobility and delivery, Sargraz Maredia, framed the next challenge in similar terms.

The next step to take in autonomous mobility is moving from individual launches to repeatable commercial scale.

Uber CEO Dara Khosrowshahi said earlier this month that the company's goal is to become the world's leading platform for commercializing autonomous vehicles, a strategy that includes gathering extensive data to share with partners and accelerate development.

Competitive landscape

The announcement comes as multiple players race to establish robotaxi services in Europe. Waymo, backed by Alphabet, operates approximately 5,000 vehicles, mostly in the United States, and has begun testing in London. In June, Waymo reportedly established new entities in four EU countries and is also planning operations in Tokyo. Chinese firms Baidu Apollo Go and WeRide are intensifying their own European plans. WeRide and Uber announced in June a pilot robotaxi project in Madrid, slated for later this year, and WeRide has separately partnered with Denmark's GreenMobility to deploy autonomous vehicles in the Nordic country in 2027. Momenta Global, another Uber partner, became the first Asian company to obtain a license to test robotaxis in Germany.

Investor pressure and regulatory headwinds

Uber has faced investor doubts over its autonomous vehicle strategy as Waymo expands rapidly. News that Waymo was exploring options to exit its existing robotaxi partnership with Uber sent shares plunging, contributing to a 20% decline from a record high in October 2025. Uber has since signed agreements with more than 30 autonomous vehicle companies. In London, Uber has secured permits with Wayve Technologies for a supervised robotaxi service and is separately testing with Baidu. For Pony.ai, the overseas push coincides with a months-long suspension of robotaxi licenses in China, triggered by a safety review after an outage of Baidu vehicles. Approvals have restarted but appear gradual, potentially slowing domestic deployment.