Pony.ai and Uber have set their sights on flooding Europe with more than 2,000 robotaxis, a move that would turn a modest Zagreb operation into a sprawling five-city autonomous fleet. The expansion plan, outlined in a Pony.ai investor notice, leaves the four additional markets unnamed for now.
The partnership leans on a three-way split of responsibilities. Pony.ai contributes Level 4 self-driving software, Uber handles bookings, payments and customer support, while local fleet operators manage day-to-day vehicle ownership and dispatch. This structure aims to bypass the costly process of building consumer networks or maintaining depots from scratch.
Yet the announcement describes a deployment target, not an operational reality. Neither company has confirmed regulatory approvals, launch dates or even which cities will host the service. Zagreb represents the only established commercial foothold, with Croatian mobility firm Verne already involved in that market.
The regulatory landscape across Europe presents formidable hurdles. Vehicle type approval exists at the EU level, but passenger services must navigate national and municipal rules covering insurance, data protection, safety reporting and remote operator oversight. The EU AI Act adds another layer where artificial intelligence functions as a safety component, though it does not alone authorize robotaxi operations.
Meanwhile, Chinese autonomous technology entering European streets will trigger scrutiny over data flows and cybersecurity. Robotaxis collect granular information about roads, passengers and surrounding activity. Regulators will demand clarity on storage locations, access controls and audit procedures for software updates. Policymakers face pressure to distinguish legitimate security requirements from blanket nationality-based restrictions.
Uber’s platform strategy allows it to partner with multiple autonomous developers rather than betting on a single in-house system. This approach reduces Uber’s technological dependence while intensifying competition among driving-system providers for access to ride demand and operational data.
For European taxi drivers, displacement risk remains uneven. Two thousand vehicles across five cities would register as significant but hardly catastrophic relative to the continent’s total private-hire market. Initial deployments may supplement existing services rather than replace them outright.
The credibility of this plan hinges on details still missing: vehicle specifications, operational domains, financing structures and evidence of granted licenses. A delivered five-city network would mark Europe’s shift from isolated trials to commercial robotaxi operation at regional scale. Without named cities and firm timelines, the announcement reads as aspiration more than imminent reality.
