Gasgoo Munich- XPENG Inc. officially rolled out its first mass-produced Robotaxi vehicle in Guangzhou on May 18, marking what the company describes as China's first Robotaxi production program developed entirely through a full-stack in-house automotive platform.
Image source: XPENG
The milestone highlights XPENG's push to vertically integrate autonomous driving hardware, software, AI models, and vehicle engineering as competition in China's Robotaxi sector intensifies.
According to an executive from XPENG's Robotaxi division, the company plans to launch demonstration operations in the second half of this year as it accelerates the commercialization of its autonomous mobility business.
During the launch event, XPENG representatives emphasized that the Robotaxi service is designed to combine premium in-cabin comfort with advanced AI-powered autonomous driving functions.
The vehicle features privacy glass, reclining zero-gravity seating, and a rear entertainment display capable of automatically deploying for passengers. Occupants can access trip information, stream entertainment content, adjust cabin settings, and interact with XPENG's in-car voice assistant, “Xiao P,” creating an experience the company says is intended to feel safer, more intelligent, and more luxurious than traditional ride-hailing services.
Built on XPENG's flagship GX platform, the Robotaxi is equipped with four in-house developed Turing AI chips delivering a combined computing power of 3,000 TOPS, which the company claims represents the highest onboard computing capability currently available in a production vehicle globally.
The vehicle also integrates XPENG's second-generation VLA large AI model, enabling Level 4 autonomous driving functions under designated operating conditions.
XPENG received regulatory approval in January to conduct intelligent connected vehicle road testing in Guangzhou, allowing the company to begin regularized Level 4 public-road testing operations.
In March, the automaker formally established a dedicated Robotaxi business unit responsible for overseeing product planning, development, testing, and operational deployment as it pushes toward commercial rollout.
Unlike many competing autonomous driving systems, XPENG's Robotaxi platform does not rely on LiDAR sensors or high-definition mapping. Instead, the company uses a pure vision-based approach combined with its second-generation VLA large model to manage real-time driving decisions.
XPENG said the architecture removes the language translation layer typically found in conventional “vision-language-action” frameworks, reducing system response latency to below 80 milliseconds. The company believes the simplified structure improves reaction speed, decision-making accuracy, and adaptability across different driving scenarios, laying the groundwork for future international expansion.
At XPENG's technology day event last year, He Xiaopeng said the company plans to open its Robotaxi SDK (software development kit) to external partners as part of a broader ecosystem strategy. Amap was announced as the first global ecosystem partner for the platform.
China's Robotaxi industry is currently entering a critical transition period as companies move from technical validation toward large-scale commercial deployment.
XPENG's vertically integrated approach — spanning vehicle engineering, AI chips, and autonomous driving software — could allow the company to accelerate the transition from development to mass operational deployment once regulatory and technical milestones are achieved.
As one of the core applications within XPENG's broader physical AI strategy, the start of Robotaxi mass production signals an important step toward scaling embodied AI technologies into real-world commercial use.
The company is positioning autonomous mobility as a potential new growth engine alongside its existing electric vehicle business, reflecting a broader ambition to expand beyond conventional passenger car manufacturing.
