Pony AI now expects to end 2026 with more than 3,500 robotaxis, a bigger target that turns its China-led rollout into a test of whether autonomous vehicles can finally scale like a real transport business.
Pony AI is no longer talking about robotaxis as a future category. It is talking about fleet size, order growth, revenue targets and vehicle costs. That matters because the autonomous driving industry has spent years selling the idea of commercial scale without always showing the operating numbers that make scale believable.
The Nasdaq and Hong Kong-listed company said on May 26 that it had lifted its year-end 2026 robotaxi fleet target to more than 3,500 vehicles, up from a previous goal of 3,000. According to Pony.ai's first-quarter earnings release, robotaxi revenue rose 395.4% year-on-year, while fare-charging revenue increased 456.5%. Average weekly paid orders in May were 119% higher than in January, and registered users in China more than tripled from a year earlier.
Those numbers are still coming from a young business. Pony AI reported total first-quarter revenue of $34.3 million, up 145% from a year earlier, but it also posted a net loss of $53.5 million. That is the central tension in the story. The market wants proof that robotaxis can move from controlled pilots to commercial networks. Pony AI is offering stronger evidence, but it is also spending heavily to get there.
Pony AI's fleet has already passed 1,700 vehicles, which gives the company a larger base from which to reach its new target. The expansion is built around its seventh-generation robotaxi, which is being rolled out across vehicles from BAIC, GAC and Toyota. The company says it is working toward a total bill of materials cost below RMB230,000 by mid-2027 for the domestic market, including both the autonomous driving kit and base vehicle.
That cost target deserves attention. Robotaxi companies do not win simply by proving that a car can drive itself. They win by proving that the car can operate all day, pick up enough riders, stay out of the repair bay, and cost less over time than a human-driven alternative. If the vehicle is too expensive, every growth target becomes a financing problem. If the vehicle gets cheaper while usage improves, the business starts to look different.
China gives Pony AI a useful operating environment for this stage. Dense cities, large ride-hailing markets and coordinated municipal support can help companies collect more miles and more service data faster than in many Western markets. Pony AI has expanded further into Guangzhou's core urban areas, including Haizhu District, Canton Tower and the Pazhou headquarters cluster. These are not empty test roads. They are the kinds of places where traffic, passengers and edge cases show up every day.
The company is also using partnerships to lower the burden of fleet expansion. Its joint deployment model with partners such as OnTime Mobility and Verne began generating revenue in the first quarter. That structure matters because the capital demands of owning and operating thousands of autonomous vehicles can quickly overwhelm even well-funded companies. Sharing fleet deployment with mobility and local partners is one way to scale without carrying every cost directly.
The U.S. comparison is getting sharper
The timing is useful because the Western robotaxi story is moving through a different phase. Waymo remains the most visible U.S. leader, but recent service pauses in Atlanta and Texas after flooding problems showed how quickly operational limits can become public issues. AP reported last week that Waymo suspended driverless service in those markets ahead of dangerous storms after vehicles were affected by heavy rain and flooding.
That does not mean Pony AI has solved everything Waymo has not. It does mean investors are now looking at two different scaling narratives. In the United States, the discussion often centers on regulatory patience, brand trust, safety cases and service reliability in difficult conditions. In China, companies such as Pony AI are increasingly pushing out larger fleet targets and trying to show that cost declines and utilization gains can carry the model.
This divergence is important for autonomous vehicle stocks. For years, AV valuations were supported by a simple belief that self-driving systems would eventually replace human drivers across large parts of transport. Now the market is asking harder questions. How many paid rides are happening? How much does each vehicle cost? Who funds the fleet? How often is the service paused? Can the system work outside the easiest weather, road and regulatory conditions?
Pony AI's answer is to lean into scale. The company now expects robotaxi revenue in 2026 to exceed 3.5 times its 2025 level, up from an earlier target of three times. It also says it remains on track to operate in more than 20 cities worldwide by year-end, with public-facing overseas services already started in Croatia, Qatar, Singapore and South Korea, and driverless deployments beginning in Dubai.
The global push could help Pony AI build credibility beyond China, but it also adds complexity. Every new market brings its own roads, regulators, consumer habits and safety expectations. A robotaxi network that works in Guangzhou does not automatically work in Zagreb, Dubai or Seoul. The advantage is that each market adds learning. The risk is that each market adds operational drag.
For now, the most important number is not just 3,500. It is whether those vehicles can generate enough rides at improving margins to make the fleet target more than a headline. Pony AI has given investors a more aggressive benchmark. The next phase is proving that a bigger robotaxi fleet can behave less like an experiment and more like a business.
Also read: Quantinuum tests public appetite for quantum computing with its IPO • Arc raises fresh funding as AI drive-thrus get another chance • Uber says AI spending needs a clearer link to useful features
