Pony.ai and Uber are putting a structure around that problem. The companies plan to collaborate on deploying more than 2,000 Pony.ai Robotaxis across Europe, expanding from Zagreb, Croatia, into four additional European cities. The cities and rollout timetable have not yet been disclosed, and the agreement also includes further deployment plans in the Middle East.
Pony.ai supplies the Level 4 autonomous-driving technology and operational expertise. Uber brings booking, payments, customer service, and demand through its mobility platform. Local fleet partners can handle day-to-day operations, while vehicle ownership and financing can sit with different participants depending on the market.
Zagreb shows how that division can work. Pony.ai, Uber, and Croatian mobility company Verne launched commercial robotaxi service there in April using Pony.ai’s Gen-7 system. Verne owns and operates the fleet and handles local market and regulatory execution; passengers can currently book through Verne, with Uber integration expected to follow.
The scale of the new commitment is notable. Pony.ai reported a global robotaxi fleet exceeding 1,700 vehicles in May 2026, meaning the planned European deployment alone is larger than that reported fleet count. The comparison is directional, however, because Pony.ai and Uber have not said how quickly the 2,000-plus European vehicles will arrive.
Pony.ai has already tested the economics behind this model in China. The company says its Gen-7 operations reached city-wide unit-economics breakeven in Guangzhou, followed by breakeven in Shenzhen in March 2026. Its joint-deployment partnerships have also begun generating revenue, giving Pony.ai a way to expand fleets with less capital concentrated on its own balance sheet. That challenge sits alongside the wider push toward turning physical AI into reliable commercial operations, where repeatability, uptime and operating economics matter more than a successful demonstration.
Pony.ai is now extending that commercialization model beyond passenger transport. In September, the company unveiled a fourth-generation Level 4 Robotruck developed with GAC Commercial Vehicle on its battery-electric T9 heavy-duty truck platform, with volume production scheduled to begin later in 2026. The technical overlap with Pony.ai’s robotaxi business is significant: the truck uses the same domain controller as its Gen-7 Robotaxis, alongside a redundant drive-by-wire architecture and an automotive-grade sensing system built from nine lidars, three millimeter-wave radars and 13 cameras. Pony.ai says the same underlying Virtual Driver technology now spans Robotaxis, heavy- and light-duty trucks, giving it a common autonomy stack across passenger and freight transportation.
The bigger test is whether that shared technology can also improve the economics of autonomous freight. Pony.ai says the bill-of-materials cost of its Gen-4 autonomous-driving kit is 70% lower than the previous generation, while it expects transportation costs per ton-kilometer to fall by 30% and energy consumption by around 10% compared with conventional trucks. Those are company projections rather than independently demonstrated fleet results, but they matter for Pony.ai’s expansion strategy: the company plans to take its Robotruck business into Europe and the Middle East over the next two years, potentially extending the partner-led model behind its robotaxi expansion into commercial logistics as well.
Europe will now test whether autonomy can become repeatable infrastructure rather than a collection of impressive city launches. Pony.ai may provide the driver, but scaling 2,000 robotaxis will depend just as much on who builds everything around it.