XPENG launches China's first mass-produced Robotaxi in Guangzhou
XPENG rolled out its first mass-produced Robotaxi in Guangzhou, marking China's first automaker-mass-produced L4 autonomous taxi.
- 3,000 TOPS
- Under 80 ms
What Happened
XPENG announced the official rollout of its first mass-produced Robotaxi in Guangzhou, making it the first automaker in China to achieve mass production of a Robotaxi through full-stack in-house development. The vehicle is built on the XPENG GX platform and engineered to L4 autonomous driving standards, powered by four self-developed Turing AI chips delivering 3,000 TOPS of computing power.
XPENG Robotaxi secured a road testing permit for intelligent connected vehicles in Guangzhou, entering routine L4 public road testing.
XPENG established its Robotaxi business unit to oversee product definition, R&D testing, and operations.
Official rollout of the first mass-produced Robotaxi in Guangzhou.
XPENG plans to initiate pilot Robotaxi operations to validate technical viability and user acceptance.
XPENG aims to achieve fully autonomous operations without on-site safety officers.
The Robotaxi uses a pure vision solution without LiDAR or high-definition maps, relying on the VLA 2.0 end-to-end large model. Its decision-making eliminates language-translation steps, compressing system response latency to under 80 milliseconds. The vehicle also features privacy glass, comfort gravity seats, and rear in-car entertainment screens controlled via voice assistant.
3,000TOPS
From four self-developed Turing AI chips, enabling L4 autonomy.
Why this matters
This rollout signals that robotaxis are moving from testing to commercial reality, with XPENG planning pilot operations later this year and aiming for fully autonomous service by early 2027.
Terms in This Story
- Robotaxi
- A self-driving taxi that operates without a human driver.
- L4 autonomous driving
- Level 4 automation where the vehicle can handle all driving tasks in specific conditions without human intervention.
- TOPS
- Trillions of Operations Per Second, a measure of AI computing performance.
- VLA 2.0
- Vision-Language-Action model that directly processes visual data into driving actions without intermediate language steps.
Summarised from the linked release; details can be imperfect — always verify against the original source.