AI Tools

Pony.ai and Uber to Deploy 2,000+ Robotaxis Across Europe

Pony.ai and Uber announced on August 13, 2026, an expansion of their partnership to deploy over 2,000 robotaxis across Europe, building on the first commercial robotaxi service in Zagreb, Croatia. The multi-city rollout will follow a joint-deployment model, with Pony.ai providing technology, Uber handling platform services, and local partners managing fleets. Each new market will require regulatory approval, with further expansion planned for the Middle East.

Neura News

Neura News

Neura Market Editorial

August 14, 20268 min read
Pony.ai and Uber to Deploy 2,000+ Robotaxis Across Europe

Pony.ai and Uber announced on August 13, 2026, that they will expand their strategic partnership to deploy more than 2,000 robotaxis across Europe. The move builds on a collaboration that began with a single commercial service in Zagreb, Croatia, and now extends to four additional, still-unnamed European cities, with further deployment planned for the Middle East.

The expansion marks a significant step for autonomous mobility in Europe, where the Zagreb service, launched on April 8, 2026, became the first fare-charging commercial robotaxi operation on the continent. The companies said rollout details will be announced in phases, and each new market will require local regulatory clearance.

From One City to a Continental Network

The partnership between Pony.ai and Uber dates back to May 2025, when the two companies first announced plans to put Pony.ai robotaxis on the Uber platform in international markets, beginning in the Middle East with safety operators aboard during an initial pilot phase. That early commitment has now evolved into a much larger ambition.

In March 2026, Pony.ai, Uber, and the Croatian mobility company Verne announced a three-way deal to launch Europe's first commercial robotaxi service in Zagreb. At that time, on-road testing of Pony.ai's seventh-generation autonomous driving system was already underway, using the Arcfox Alpha T5 robotaxi developed jointly by Pony.ai and the Chinese automaker BAIC.

The Zagreb service went live on April 8, 2026, with the public booking paid rides through the Verne app. The initial zone covered roughly 90 square kilometers, including central Zagreb and the city's airport, and operated daily from 7:00 a.m. to 9:00 p.m. Integration into the Uber app was slated to follow, and the August expansion announcement confirms it is still coming.

The expanded agreement now commits to deploying more than 2,000 Pony.ai robotaxis across Europe, extending well beyond Zagreb. The four additional European cities have not been named, and the companies said rollout details will be announced in phases. The Middle East, which was part of the original May 2025 announcement, remains in the plan.

A Three-Part Model for Scale

The expansion is structured around Pony.ai's joint-deployment model, which splits three functions across specialized partners. Pony.ai supplies the Level 4 autonomous driving technology and operational expertise drawn from its existing large-scale deployments. Uber provides booking, payment, and customer service through its platform. Fleet operations may be run by established local partners, as Verne does in Zagreb.

This structure is designed to keep capital burden off any single balance sheet. Vehicle funding and ownership can sit with different partners depending on the market, and the expanded agreement keeps the option of established local fleet partners selected market by market.

Dr. James Peng, founder and CEO of Pony.ai, said, "This expanded agreement marks an important new phase in the partnership between Pony.ai and Uber," adding, "By combining Pony.ai's proven autonomous driving technology and operational know-how with Uber's global mobility platform and extensive market reach, we aim to build sustained commercial operations at scale across Europe and beyond."

Uber's role in the partnership remains consistent: customer access, not vehicles. The company does not develop self-driving technology in-house and instead maintains a roster of autonomous partners. Sarfraz Maredia, Global Head of Autonomous Mobility & Delivery at Uber, said, "The next chapter for autonomous mobility is about moving from individual launches to repeatable commercial scale," a line that captures the shift from the single Zagreb operation to a multi-city network.

The Zagreb Template

The Zagreb rollout provides a clear template for what may come in the four additional European cities. The sequence runs from public-road validation first, then a fare-charging launch with a named local operator, and finally platform integration.

In Zagreb, Verne led market readiness and European regulatory approval, owning and operating the robotaxi fleet. Uber committed to investing in Verne as part of the March 2026 three-way deal. The service operated daily from 7:00 a.m. to 9:00 p.m., covering central Zagreb and the airport within its roughly 90-square-kilometer zone.

The #1 Newsletter in AI

Stay ahead of the AI curve

The most important updates, news, and content — delivered weekly.

No spam. Unsubscribe anytime.

The August expansion announcement confirms that integration into the Uber app in Zagreb is still coming, which would complete the template in that city. If the four-city rollout proceeds on the same pattern, each new market would begin with public-road testing, then move to a fare-charging service with a local operator, and finally connect to the Uber platform.

Each new market will require local regulatory clearance, and the companies have not specified which cities are next or when they might launch. The expanded agreement keeps the option of established local fleet partners selected market by market, suggesting that Verne-style operators could emerge in other European cities.

Pony.ai's Track Record and Technology

Pony.ai, founded in 2016 and based in Guangzhou, is listed on both NASDAQ and the Hong Kong Stock Exchange. The company runs paid, fully driverless robotaxi services in China's four tier-one cities and reports city-wide breakeven unit economics in multiple markets in China, a company claim that underpins its push into Europe.

The company's Gen-7 system, which powers the Arcfox Alpha T5 robotaxi used in Zagreb testing, is described by Pony.ai as enabling mass production of robotaxis at lower cost. As of earlier in 2026, Pony.ai was targeting a fleet of 3,000 vehicles by year-end, and the European commitment of more than 2,000 robotaxis would substantially extend that figure.

The technology has already proven itself in commercial service. The Zagreb operation, which went live on April 8, 2026, made Verne's operation the first fare-charging commercial robotaxi service in Europe. Public riders booked paid rides through the Verne app from day one, and the service has been running daily since.

Capital and Competition in the Autonomous Space

The expansion comes as the autonomous vehicle sector attracts significant capital. Moove, a fleet operator, raised a $250 million Series C at a $2.1 billion valuation to scale robotaxi fleet infrastructure, an example of the fleet-financing capital flowing into the ecosystem. The joint-deployment model that Pony.ai and Uber are using is designed to keep the capital burden off any single balance sheet, which may make it easier to scale across multiple markets.

Uber's autonomy strategy is deliberately diversified. The company does not develop self-driving technology in-house and instead works with a roster of autonomous partners. This approach allows Uber to focus on customer access while partners handle the vehicles and technology. In the US market, the pattern involves established operators moving from waitlists to open commercial service city by city, as Waymo did in Dallas.

The European expansion represents a different path, one built on the Zagreb template. The bet is that a single 90-square-kilometer service zone can become a continental network of more than 2,000 vehicles. The companies have not said when the four additional cities will be named or when the Middle East deployments will begin, but the phased rollout suggests a gradual approach.

The expanded agreement extends the Middle East region that the partnership has targeted since its first announcement in May 2025. That original plan called for safety operators aboard during the initial pilot phase, and the new agreement keeps the region in scope. The companies said rollout details will be announced in phases, leaving the timing and specific locations open.

For now, the focus remains on Europe, where the Zagreb service has established a working model. The four additional cities, still unnamed, will test whether that model can be repeated at scale. Each will require local regulatory clearance, and each will need a local operator if the Zagreb template holds. The companies have not said which cities are next, but the commitment to more than 2,000 robotaxis across Europe signals a substantial build-out.

Pony.ai's Gen-7 system is central to that build-out. The company describes it as enabling mass production of robotaxis at lower cost, which would be essential for deploying more than 2,000 vehicles across multiple markets. The Arcfox Alpha T5, developed with BAIC, is the vehicle platform for the system, and it has already completed on-road testing in Zagreb.

The partnership between Pony.ai and Uber has moved quickly from announcement to operation. The May 2025 announcement of international plans, the March 2026 three-way deal for Zagreb, the April 8, 2026 launch, and now the August 13, 2026 expansion cover a span of just over a year. The next phase, naming the four additional cities and beginning Middle East deployments, will determine whether the model scales beyond a single city.

Related on Neura Market

More from Neura News

Research

LittleLearner Models Trained Only on K-5 Curriculum Show Skills Are Elicited, Not Acquired

Researchers released LittleLearner, a family of language models trained from scratch on a strictly filtered K-5 elementary school curriculum, to answer whether capabilities beyond training data can be elicited or acquired through scaling, post-training, and in-context learning. The answer is largely no: scaling, post-training, and in-context learning amplify what the curriculum taught, but none meaningfully improve out-of-scope performance. The pretraining filter sets the effective capability ceiling, providing a controlled sandbox for studying knowledge acquisition and RL.

Aug 16·5 min read
Industry

The Hidden Gold Rush: Scammers Exploit Demand for Claude Watermark Removal Apps

Anthropic's August 2026 watermarking of Claude text has sparked a surge in demand for removal apps, attracting scammers who peddle fraudulent tools. AI scientist Lance Eliot warns these apps often contain malware or fail to work, as statistical watermarks are nearly impossible to remove without heavy editing. With billions of users at risk, the problem is expected to worsen as more AI makers adopt watermarking.

Aug 16·12 min read