China robotaxis top 1,100 vehicles

Robotaxis in China are no longer a science-project story. They are turning into a fleet-operations story. That sounds less glamorous, but it is the part that decides whether autonomous driving becomes a real transport business or just a really good demo. The change this year is that the big Chinese players are adding vehicles, adding cities, and just as important, building the boring support systems that keep those vehicles working. ### Which companies are actually moving? Three names keep showing up — Pony.ai, Baidu’s Apollo Go, and WeRide. Pony.ai has been especially aggressive in 2026, laying out a plan for more than 3,000 autonomous vehicles across 20-plus cities and tying that expansion to mass-produced seventh-generation robotaxis. WeRide is already operating at meaningful fleet size, saying it had more than 1,200 vehicles in service as early as the first quarter of 2025 and has kept expanding internationally. (sec.gov) Baidu is still the volume benchmark on rides, with Apollo Go reporting more than 17 million rides globally by the end of October 2025. ### Why does the 1,100-vehicle number matter? Because it points to density, not just bragging rights. A fleet can look big on paper while still being scattered across pilots, showcase routes, and low-usage test zones. Once you start seeing four-digit totals across active deployments, the question changes from “can the car drive itself?” to “can the operator keep enough cars charged, cleaned, routed, and legally deployed to make money?” That is a much harder version of the problem — and a much more commercial one. (sec.gov) The available disclosures from Pony.ai, WeRide, and Apollo Go suggest that China-linked operators are now firmly in that phase. ### Why are Dubai and Singapore part of this? Because Chinese robotaxi companies are not staying inside China anymore. Pony.ai deployed driverless robotaxis in Dubai in April and said it expects public fare-charging service there later in 2026 with a fleet in the hundreds. The same company also moved its Singapore program into by-invite rides with ComfortDelGro — basically the last stop before public-facing service. WeRide launched fully driverless fare-charging robotaxi operations in Dubai in March 2026 and already had a strong Abu Dhabi footprint before that. (sec.gov) ### What is the real bottleneck now? Operations. A robotaxi company does not win just by having better autonomy software. It also needs depots, charging, maintenance, dispatch, permits, and enough utilization to spread fixed costs over a large fleet. That is why the launch of Neolix and TELD’s “Power Island” matters more than it first appears. The Qingdao site is pitched as an autonomous charging and operations center, with one location able to charge 90 vehicles per hour and handle parking, cleaning, diagnostics, and maintenance without human plug-in work. (blog.pony.ai) ### Where does Chery fit in? Not mainly in robotaxis. Chery’s AiMOGA unit is showing how adjacent autonomy hardware is scaling too. At its April 2026 event, the company highlighted contracts for 1,000 intelligent police robots and delivery of 110 units. That is not the same market as ride-hailing, but it shows the same broader pattern — Chinese autonomy companies are trying to industrialize fleets, not just prototype machines. (pandaily.com) ### Are these businesses actually growing? Yes, but unevenly. Pony.ai told investors it scaled top-line growth, fleet size, and footprint in 2025, and said multiple tier-one Chinese cities reached unit-economics breakeven. WeRide reported full-year 2025 revenue up 90%, with robotaxi revenue up 66.4% in the fourth quarter. Those are still early-stage numbers, but they matter because they hint the business is moving closer to repeatable economics instead of permanent subsidy mode. (news.marketersmedia.com) ### So what changed this year? The center of gravity moved from autonomy as a tech feat to autonomy as a logistics system. The cars still matter. The software still matters. But now the winning question is whether Chinese operators can run large, regulated, charged, maintained fleets across multiple cities and countries without the economics falling apart. That is what makes the current buildout feel different. ### Bottom line? China’s robotaxi story is getting bigger, but more important, it is getting more boring in the right way. (sec.gov) Once the headlines shift to depots, permits, utilization, and charging islands, that usually means the industry is finally trying to become real.