Pony.ai now has more robotaxis promised to partners outside China than it has driverless vehicles on the road anywhere in the world.
The Guangzhou-based autonomous driving company disclosed Tuesday, alongside its second-quarter results, that its pipeline of planned and potential overseas robotaxi deployments has expanded to more than 4,000 vehicles. For scale: Pony.ai’s entire global fleet stood at 1,975 vehicles as of June 30. The international order book is now roughly double everything the company operates today, China included.
The anchor is Uber. Pony.ai has a contracted deployment of more than 2,000 robotaxis across five European cities with the ride-hailing giant — a deal management says makes it Uber’s largest autonomous driving partner in Europe. The balance of the 4,000-plus figure covers agreements in negotiation with partners including Bolt, which is working alongside Pony.ai and Stellantis in Luxembourg, and Singapore’s ComfortDelGro, whose Zig app opened Pony.ai robotaxi rides to the general public during the quarter.
“With this contract, we become Uber’s largest autonomous driving partner in Europe,” chairman and chief executive James Peng told analysts on the earnings call. “Going forward, as the performance and also the economics continue to validate at scale, we’ll see substantial room to expand the fleet size even further.”
Peng framed the win as vindication. When the two companies first worked together in early 2025, he said, “there were some doubts whether our autonomous driving capabilities can handle the European cities.” Those doubts, he argued, have since been “answered with resounding real-world evidences.”
The numbers behind the push
The pipeline disclosure landed on top of a quarter that gave the international story financial credibility. Total revenue rose 68.8% year over year to US$36.2 million. Robotaxi services revenue hit US$12.1 million, up 691.2% — an acceleration from the 395% growth posted in the first quarter — with fare-charging revenue, the purest measure of paid rides, up 849.3%.
Gross margin improved to 17.5% from 16.1%, and the loss picture narrowed considerably in relative terms: operating loss margin fell to 181.5% from 285.6% a year earlier, and net loss declined 14.9% to US$45.4 million. Pony.ai ended June with US$1.39 billion in cash and investments, against US$32.2 million of capital expenditure in the quarter.
The mechanism that makes 4,000 overseas vehicles plausible on that balance sheet is what Pony.ai calls its joint deployment model: the company supplies the AI driver and its seventh-generation robotaxi hardware, a mobility platform such as Uber or Bolt supplies demand, and a local operator handles fleet management and maintenance. Crucially, the partners fund the fleet.
“This model could generate sharing-based revenue or technology licensing fee for Pony,” chief financial officer Leo Wang said. “This is not only broaden our revenue base, but also introduce higher margin recurring income across the entire Robotaxi operating life cycle.”
The other enabler is software. Chief technology officer Tiancheng Lou credited PonyWorld 2.0, the company’s world model, with letting engineering headcount stay flat as city count rises — including at the launch in Zagreb, Croatia, which Pony.ai describes as the first commercial robotaxi service in a European capital’s city center. “This turns cities expansion from effort that used to take dozens of engineers into a human-in-the-loop automatic process that just a few people can run,” Lou said. Fleet operations now run at roughly three staff per 100 robotaxis.
Why it matters
The 4,000 number is best read as a competitive signal rather than a delivery schedule. These are commitments and agreements in negotiation, and go-live dates depend on permitting and regulatory clearance in each jurisdiction. Nothing about a signed contract guarantees a vehicle carrying passengers in Munich or Riyadh next spring.
What it does establish is the shape of the global robotaxi market forming in 2026. Waymo has scale and regulatory trust in the United States but essentially no international footprint. Chinese operators — Pony.ai, Baidu’s Apollo Go, and WeRide — have responded by racing into the markets Waymo has not claimed: the Gulf, Europe, Southeast Asia. Apollo Go has begun public road testing in London with Uber and Lyft and expanded into South Korea; WeRide, operating with Uber in Abu Dhabi and Dubai, is targeting a global fleet of roughly 2,600 vehicles by year-end. Uber has quietly become the kingmaker, routing demand to whichever AI driver clears its safety and cost bar in a given city.
That reframes the AI-infrastructure debate. Robotaxis are the most capital-intensive consumer deployment of foundation-model-style AI anywhere, and the winner will likely be decided less by raw model quality than by who can amortize a driving stack across the most jurisdictions. Pony.ai’s pitch is that a world model plus an asset-light partner structure lets it enter a market without rebuilding the stack — a claim Lou put bluntly when analysts raised Waymo’s remark that a demo represents only 1% of the work. “If all you have today is a demo, you still need to prove that you can achieve multiple 10x performance jumps,” he said. “On top of that, you need time to build trust with regulators before you can scale.”
The counterweights are real and mostly non-technical: European and Gulf regulators approve deployments city by city; insurance frameworks for driverless liability remain immature; and a Chinese company mapping and recording foreign streets invites data-security scrutiny that has already complicated Chinese technology exports elsewhere.
What to watch next
Three markers over the next two quarters. First, conversion: how much of the 4,000-vehicle pipeline moves from agreements in negotiation to signed, permitted deployments — and how fast the first Uber-partnered European cities go live. Second, the year-end targets, which Pony.ai reaffirmed at more than 3,500 vehicles and operations in more than 20 cities, plus confidence in beating its original robotaxi revenue guidance of exceeding 3.5 times last year’s level. Third, margins. Non-GAAP operating expenses rose just 9.6% against 68.8% revenue growth this quarter. If that gap holds while the fleet doubles, Pony.ai will have shown that scaling driverless mobility internationally does not require scaling cost with it. If it does not, 4,000 stays a number in a slide deck.
“With this contract, we become Uber's largest autonomous driving partner in Europe. Going forward, as the performance and also the economics continue to validate at scale, we'll see substantial room to expand the fleet size even further.”— James Peng, Chairman and CEO, Pony.ai