Uber's partner-fleet plan for 2,000 European robotaxis goes to London for testing

Spread evenly, more than 2,000 robotaxis across five European cities comes to about 400 cars a city ^[2][17]. Wayve, the British startup already carrying Uber passengers in London, runs 15 Ford Mustang Mach-Es with safety drivers aboard, according to Crypto Briefing ^[5]. One average city in the Uber and Pony.ai plan is roughly 27 times that fleet ^[18].

Rider interest is not the scarce input. Crypto Briefing reported that more than 140,000 London users have opted in to Wayve's service ^[9], about 9,300 for each of its 15 cars ^[19]. An opt-in is not a paid trip. Still, on these numbers the shortage is cars, and someone to buy them.

Zagreb, where TechCrunch says the 2,000-car rollout starts ^[3], handled that with a third company. Pony.ai supplies the driving system and an Arcfox Alpha T5 developed with BAIC, Verne owns and operates the fleet, and Uber brings its ride-hailing network, TechCrunch reported ^[4]. Uber did not build the driving software ^[7]. In Zagreb it does not own the cars either ^[4]. TechCrunch said London will follow a similar model, though the companies have not said who will own and operate the fleet there ^[8].

If a Verne-type owner is named for London, the Zagreb split repeats, and on that template the European plan needs a fleet financier in each of its five cities ^[2][4]. The alternative is Uber or Pony.ai taking the London cars onto its own books. That would put vehicle costs on one of the two companies that left ownership to Verne in Zagreb ^[4]. Transport for London is the third variable: Crypto Briefing reports that the city's transport authority holds real influence over street operations on top of the national Automated Vehicles Act ^[15]. Slow approvals there would push the 2,000 cars toward easier cities first.

I think the ownership announcement will say more about whether the European plan can be built than the London road tests will. The case against that view is Crypto Briefing's own description of London as one of the harder places to teach a car to drive, with narrow streets, heavy foot traffic and road layouts older than the combustion engine ^[16]. If Gen-7 cars cannot cope with those streets, nobody needs to finance a London fleet. The view is also wrong if Uber ends up owning the London cars, because its partner model would then carry vehicle costs on Uber's own books in at least one European city.

The supplier count fits London getting two partners at once. Uber has struck deals with more than 30 autonomous-vehicle companies ^[11] and expects AV trips in up to 15 cities by the end of 2026 ^[12]. That is about two suppliers for every target city ^[21]. Dubai already has Baidu and WeRide on separate Uber deals ^[13]. In London, Baidu's Apollo Go and Waymo are also exploring operations, according to Crypto Briefing ^[14].

Wayve is also an Uber investment with a condition attached. Wayve raised $1.2 billion in February 2026 from investors including Microsoft, Nvidia and Uber. TechCrunch reported that the round could reach $1.5 billion with an extra $300 million from Uber, contingent on Wayve deploying robotaxis, starting in London ^[10]. The contingent tranche is 20% of the maximum ^[20]. The two reports disagree on where Wayve stands. Crypto Briefing describes its supervised service as live since early September ^[5], while TechCrunch describes Wayve as planning a London launch ^[6]. Uber is bringing Pony.ai into the same city where, according to TechCrunch, $300 million of its Wayve commitment depends on deployment ^[10].