China's robotaxi sector is accelerating its shift toward scaled commercial operations, with leading companies expanding fleets in major cities, as technology advances and regulators refine rules for road use.

Chinese autonomous driving companies Pony.ai and WeRide said their robotaxi fleets have each surpassed 1,000 vehicles, reaching 1,159 and 1,023, respectively.

The 1,000-vehicle mark is widely seen as a key threshold, signaling a shift from technology validation to routine service, analysts said.

Pony.ai has deployed its robotaxi service in Beijing and Shanghai, as well as Guangzhou and Shenzhen in Guangdong province, all of China's four top-tier cities. In late 2025, the company announced that its robotaxi fleet in Guangzhou had achieved per-vehicle profitability. The development is seen as a step toward commercially sustainable driverless ride-hailing services.

In addition, Pony.ai has set a 2026 target to expand its overall robotaxi fleet to more than 3,000 vehicles. Wang Haojun, CFO of Pony.ai, said the next stage is about increasing volume and stepping up deployments.

He added that robotaxi operations ultimately depend on scale effects.

WeRide said it has expanded its robotaxi business globally, with a presence in more than 30 cities. It began fully driverless robotaxi commercial operations in Abu Dhabi, the United Arab Emirates, in late November. In Abu Dhabi, each vehicle averages about 15 to 20 ride requests a day, and profitability is expected once the fleet reaches around 200 vehicles.

Apollo Go, Baidu's autonomous ride-hailing service, said its robotaxi operations in China now cover about 20 cities. On Jan 17, the company said it had moved forward with AutoGo, a UAE autonomous mobility company. The two sides have started fully driverless commercial robotaxi operations in Abu Dhabi.

These developments have been underpinned by faster technology iteration and policy measures. Specifically, advances in software and lower hardware costs are supporting robotaxi fleet expansion. Techniques such as bird's-eye view perception, occupancy networks and end-to-end large models are being used to improve performance in urban traffic.

Pony.ai said the materials cost of its seventh-generation autonomous driving kit has fallen by nearly 70 percent.

WeRide also said breakthroughs in technology have reduced its autonomous driving kit cost by 50 percent and lowered total cost over its full life cycle by 84 percent, making larger fleet rollouts more viable.

Policy and standards work is helping speed up large-scale deployment by providing a clearer regulatory framework.

The Ministry of Industry and Information Technology has urged faster progress on standards for autonomous driving, including rules on the conditions and scenarios in which autonomous systems can operate and on simulation testing. It also called for mandatory national safety standards for autonomous driving systems.

Local governments are refining compliance rules for autonomous driving deployment. Shenzhen, for example, has introduced regulations on intelligent and connected vehicles, outlining a preliminary legal framework for issues such as accident liability and data security.

Meanwhile, Shanghai revealed a plan to advance autonomous driving on Jan 14. It set a 2027 target to scale Level 4 applications in services including robotaxis and robobuses. It aims for more than 6 million passenger trips, open testing areas of about 2,000 square kilometers and road coverage over 5,000 kilometers in the year.

At the national level, China has opened more than 35,000 km of test roads and selected about 20 cities for vehicle-road-cloud integration pilots to support the large-scale deployment of high-level automated driving.