Uber and Pony.ai Bring Robotaxis to London

Uber and Pony.ai plan to launch a robotaxi service in London, extending their partnership to bring driverless cars to Europe. Testing of Pony.ai's Gen-7 robotaxis, the company's seventh-generation driverless model, is scheduled to begin in the coming weeks TechCrunch.

The London deployment is tied to a wider rollout target. Uber expects to offer autonomous vehicle trips, rides with no human driver, in as many as 15 cities globally by the end of 2026. London would be one of the hardest operating environments in that set, with dense mixed traffic, irregular road layouts and close interaction with buses, cyclists and black cabs.

Uber and Pony.ai first partnered in May 2025, initially focusing on the Middle East. The European phase centers on a plan to deploy 2,000 robotaxis, starting with Zagreb. In August 2026, Uber, Pony.ai and Verne launched an autonomous ride-hailing service in that city, giving the partners an operating reference before attempting London.

Fleet scale matters here. Pony.ai, a Chinese developer of automated driving systems, reported its robotaxi fleet had scaled to over 1,400 vehicles as of March 25, 2026. That provides baseline capacity for parallel deployments, though vehicles assigned to one city cannot simply move to another without local mapping, testing and regulatory clearance.

London is not a single-vendor effort for Uber. In March 2026, Wayve, Uber and Nissan announced a collaboration on robotaxis Uber Investor. Wayve and Uber described a global rollout across more than 10 cities, including London, plus a pilot in Tokyo for late 2026. Uber had said on June 8, 2026 that London's first robotaxis were expected within months, using AI self-driving technology from British startup Wayve. Users could sign up for a chance to ride in those early vehicles.

A separate deployment shows how Uber is fitting third-party AV stacks, outside self-driving software plus sensors, into its dispatch and payments layer. WeRide and Uber launched autonomous robotaxi rides in Dubai, expanding Uber's autonomous footprint in the UAE Uber Investor. Riders can book a WeRide robotaxi through the Autonomous option in the Uber app in select locations. The model is consistent across partners. Uber supplies demand aggregation, routing and customer interface. The AV partner supplies the driving stack, vehicle platform and fleet supervision.

The broader context here is that Uber is pursuing a multi-stack, multi-OEM strategy rather than betting on one driving system. In simple terms, it is using self-driving systems from several firms and vehicles from several makers. Pony.ai, Wayve with Nissan, WeRide and Verne each bring different sensor setups, test histories and operating areas. That spreads technical risk and gives Uber leverage in talks. It also increases integration load, since each stack needs separate safety cases, remote-support workflows, maintenance plans and city approvals.

In my view, London will test that strategy more strictly than Dubai or Zagreb. Scale is only part of the problem. Unusual road cases are more frequent, weather degrades perception more often, and public scrutiny is intense. Success will depend less on headline vehicle counts and more on mean time between interventions, or how long cars run before needing human help, depot throughput and the ability to keep use high without lengthening pickup times.

What is worth watching longer term is what this enables if execution holds. A single app that can dispatch human drivers, Wayve-driven Nissans and Pony.ai Gen-7 vehicles in the same city turns autonomy into a routing problem rather than a separate product. That makes mixed fleets workable during the long shift when neither human-only nor autonomy-only supply is enough. Technology has absorbed this kind of complexity before to improve daily life, and urban transport is a likely place for it to do so again.