Pony.ai paid rides surge sixfold

Pony.ai just dropped a useful datapoint for the robotaxi business — not a demo, not a pilot, but a burst of real paid demand during China’s May 1–5 Labor Day holiday. The company said average daily paid robotaxi orders jumped 544% from a year earlier and 155% from the New Year holiday. That matters because the big unanswered question in autonomy has never been just “can the car drive?” It’s “will normal people actually use this enough to build a business?” Pony.ai is arguing that the answer is starting to look like yes. (prnewswire.com) ### Why is a holiday spike such a big deal? Holiday traffic is messy in exactly the way robotaxi operators need to prove they can handle. Demand comes in waves. Pickup zones get chaotic. Riders are often unfamiliar with the area. And the trip mix shifts from routine commuting to event traffic, tourism, and late-night surges. If a service holds up there, that says more than a quiet weekday loop in a geofenced district. Pony.ai tied the jump to both fleet expansion and broader consumer adoption, which is basically the whole commercialization story in one sentence. (prnewswire.com) ### What actually happened in Guangzhou? The cleanest proof point was a concert run in Guangzhou’s Nansha district. A Teens in Times concert series pulled about 220,000 fans and family members across four shows, with more than 95% coming from outside the city. Pony.ai said it sent hundreds of robotaxis into the post-show rush from 10 p.m. to 2 a.m. over four consecutive nights and moved more than 3,000 concertgoers. Even more interesting — the service was included in official post-event transportation guidance. That means the robotaxis were not treated like a novelty side option. They were part of the traffic plan. (prnewswire.com) ### Is this just one good weekend? Not really — the bigger pattern started earlier. Pony.ai’s March 2026 annual-report update said its robotaxi business hit city-wide unit-economics breakeven in Guangzhou in November 2025 and in Shenzhen in February 2026. It also said paid orders in Shenzhen from January to mid-February 2026 already topped the city’s total for all of 2025, and a peak day reached 25 orders per Gen-7 vehicle with daily net revenue of RMB394. So the Labor Day jump looks less like a fluke and more like an acceleration on top of an improving base. (www1.hkexnews.hk) ### Why does fleet size matter so much? Because robotaxi demand is weirdly hard to read when supply is constrained. If wait times are long or coverage is patchy, usage numbers understate real demand. Pony.ai says its active robotaxi fleet reached 1,446 vehicles by March 25, 2026, and it still plans to get past 3,000 by the end of 2026, with service in more than 20 cities globally and nearly half the fleet overseas. In other words, the company is trying to answer the chicken-and-egg problem by brute force — put enough cars on the road that rider behavior stops being distorted by scarcity. (sahmcapital.com) ### So is China now the easiest place to scale robotaxis? Easier, yes. Frictionless, no. Pony.ai has a real edge because it already holds fully driverless commercial permits across Beijing, Shanghai, Guangzhou, and Shenzhen — all four of China’s tier-one cities. That is unusually broad regulatory coverage. But the catch is that regulation can still tighten fast. Reuters reported in late April that China had paused new autonomous-vehicle permits after a Baidu Apollo Go incident in Wuhan. So the market is opening, but it is still permissioned growth, not software-style growth. (ir.pony.ai) ### What’s the real takeaway? The important part is not the holiday headline by itself. It’s that paid usage, unit economics, fleet scale, and city-level permissions are finally starting to line up at the same time. That does not mean robotaxis are solved. But it does mean Pony.ai is moving from “autonomy company with a future market” toward “transport operator with measurable demand” — and that is a much more serious place to be.