The EV parts bin behind China's driverless taxi push
Baidu, WeRide, and Pony.ai already run paid driverless taxis in Beijing's Yizhuang district, and China's edge is the EV supply chain they inherited.
Baidu, WeRide, and Pony.ai already run paid driverless taxis in Beijing's Yizhuang district, and China's edge is the EV supply chain they inherited.
Driverless taxis already pick up paying passengers in Beijing's Yizhuang district, a commercial robotaxi zone where Baidu, WeRide, and Pony.ai run overlapping services. The scene looks like a breakthrough. The structural story is industrial. China's robotaxi industry reuses the batteries, sensors, motors, and assembly lines that already made China the world's largest EV market, and that inheritance now compresses the cost of autonomous driving in ways the rest of the race will have to match.
Robotaxis need lidar, radar, cameras, high-voltage wiring, electric drive units, and large onboard compute, and every one of those is already in volume production for Chinese passenger EVs from suppliers such as BYD, Geely, Chery, and SAIC. Baidu's Apollo Go, Pony.ai, and WeRide avoid building a parallel automotive supply chain because they plug into one. Market analysts have framed the listed trio, BIDU, PONY, and WRD, as the primary investable expression of that inheritance because each carries the operator's stack while drawing on a shared industrial base.
Yizhuang makes the inheritance visible on the road. The district mixes buses, cyclists, scooters, and pedestrians, and the robotaxis navigate with minimal remote intervention. Riders book with an app. Operating at this density, in mixed traffic, with commercial services, is the part of the robotaxi story that has historically separated demos from businesses, and reaching that point first in a Chinese pilot zone follows from which country already had the parts and software base in volume.
By mid-2025, Chinese-headquartered robotaxi firms were already operating commercial or pre-commercial services in the Middle East, ahead of equivalent US deployments in those markets, and multiple operators have pointed to commercial viability milestones in 2026. The export pattern matters because it treats foreign markets as a near-term extension of an already-built domestic platform, while a US firm entering the same geography has to do the platform build first.
The US counterweight is real. Waymo is scaling robotaxi service in multiple US cities, and Tesla has signaled plans for its own Cybercab service. American companies have their own supply chain advantages in compute and foundation models. The competitive question shifts from who has the better sensor kit on a slide to who can field a robotaxi service at a cost and cadence the existing auto industry is already set up to support, and the Chinese side entered that race with the world's largest EV factory floor already paid for. Industry analysts frame this as a near-term showdown between Chinese operators and the US incumbents.
The caveats carry weight. Yizhuang is a designated testing ground, not a national permission slip, and public-road rollout elsewhere still depends on local regulators, safety incident volume, and the technology's ability to handle unstructured environments the training data does not cover. Some of the bullish framing on the robotaxi race traces back to investment commentary rather than independent reporting, and the technology, while past the demo stage, has not yet had to absorb the cost of a large fleet-wide incident.
The concrete watch item is 2026. Baidu, Pony.ai, and WeRide are targeting commercial viability milestones this year, from per-trip economics to fleet counts, and the Middle East push gives analysts a way to score progress outside the pilot zones. If the milestones land, the test for US competitors is whether they can field an equivalent platform before the cost advantage compounds further.