XPeng launches robotaxi service in Guangzhou, trailing leaders by 3 years

XPeng officially began production of its first purpose-built robotaxi in Guangzhou on Monday, marking a carefully measured entry into China’s increasingly competitive autonomous ride-hailing market. The company frames this as a milestone: the first Chinese automaker to mass-produce a robotaxi developed entirely with proprietary technology.

The vehicle rides on XPeng’s new GX platform and is engineered from the ground up for Level 4 autonomous driving. Unlike most robotaxis currently in operation, which are retrofitted consumer cars, this model integrates the necessary compute, drive-by-wire chassis, and redundancy directly at the factory level.

That distinction matters. XPeng is not the first company in China to operate robotaxis commercially. Baidu’s Apollo Go runs over 1,000 vehicles across 22 cities, making it the country’s largest operator. Pony.ai had 1,446 vehicles in its fleet by late March and aims for 3,000 by year-end. WeRide operates more than 1,000 units across 30 cities in 11 countries. What XPeng can legitimately claim is being the first traditional automaker, rather than a pure-play autonomous driving company, to put a robotaxi-grade vehicle into series production in China.

That distinction blends marketing with real substance. Until now, pure-play operators have retrofitted consumer EVs, often Toyotas or domestic models, with sensor arrays and onboard computers supplied by the AV company. XPeng’s approach is the reverse: a vehicle designed from the platform up with the hardware and software required for Level 4 operation.

The robotaxi carries four of XPeng’s own Turing AI chips, delivering a combined 3,000 TOPS of compute power. It uses a Bosch next-generation steer-by-wire system that eliminates the mechanical steering shaft, and XPeng says it has built aviation-grade redundancy into all safety-critical systems. The software stack, called VLA 2.0, is a vision-language-action model that compresses end-to-end response latency to under 80 milliseconds.

Production volume will start small. Brian Gu, XPeng’s president, told Reuters last month that the company expects to produce hundreds to thousands of robotaxis over the next 12 to 18 months. Pilot operations with safety drivers begin in the second half of this year, and fully driverless commercial service is targeted for early 2027. The Guangzhou production line already holds an intelligent-connected-vehicle road-testing licence, and L4 road tests are running on public roads.

This robotaxi push sits inside a broader diversification strategy for 2026. XPeng is also developing humanoid robots and a modular flying car, all powered by the same Turing silicon and vision-language-action software stack. The company has been clear that it wants to monetize its AI stack across multiple form factors rather than rely solely on EV margins in a market where prices are falling and government subsidies have been reduced.

Volkswagen is the most visible external customer for that strategy. The German automaker, which invested $700 million for a 4.99% stake in XPeng in July 2023, named XPeng’s Turing chip and VLA 2.0 system as its first external commercial customer this year. It is the first time a major Western carmaker has adopted Chinese-developed autonomous driving software at this depth. XPeng has indicated it targets roughly one million Turing chip shipments in 2026 and plans to open its robotaxi SDK to additional fleet operators over time.

The timing of Monday’s announcement is complicated by a softer near-term outlook. XPeng guided first-quarter 2026 deliveries to 61,000–66,000 units, a year-on-year decline of roughly 30% that the company attributes to reduced government subsidies and a longer Lunar New Year. The company reports Q1 earnings on Wednesday. Its first-ever quarterly net profit of RMB 380 million, posted for Q4 2025, arrived alongside that cautious Q1 guide.

The operational risks of large autonomous fleets have also become more visible. In late March, more than 100 Baidu Apollo Go robotaxis froze mid-traffic in Wuhan due to a correlated software failure, stranding passengers and surfacing a category of risk that conventional vehicle regulation was not designed for: simultaneous failure across a managed fleet. Pony.ai, despite reaching city-level profitability in Guangzhou, faces similar concerns as it scales.

XPeng enters the same operating environment with a hardware advantage and a software stack it has built and tested itself. But it also enters as a software operator, not just a manufacturer, with the same governance questions about fleet-scale failure modes that existing players are still working through.

What XPeng has done this week is start the line. The first car has been built. The second-half pilots will test whether the platform performs in mixed traffic with a safety driver on board, and the early-2027 target will test whether the company can remove that driver. Hundreds to thousands of units over 18 months is a deliberately modest ramp by the standards of Baidu and Pony, both of which target fleets in the low thousands by the end of this year.

XPeng is betting that integration,owning the chip, the software, the platform, and the chassis,will produce a more reliable robotaxi than the retrofitted models that currently dominate the market.