newsio aggregates and links to original sources. We do not own the original images or content. If you believe content infringes on intellectual property rights, contact us — it will be removed at first notice.
robotics/news//Interesting Engineering
XPENG has rolled out China’s first production-ready, pre-assembled robotaxi developed entirely in-house.
XPENG has developed China's first production-ready, fully in-house robotaxi built on the GX platform.
KEY POINTS
The Robotaxi operates entirely without LiDAR or high-definition maps, using a pure vision VLA 2.0 model.
XPENG's proprietary Turing AI chips deliver 3,000 TOPS, eliminating reliance on Nvidia silicon.
XPENG targets fully driverless commercial Robotaxi operations, without safety officers, by early 2027.
The VLA 2.0 large model also powers XPENG's humanoid robot IRON and its flying car.
XPENG has rolled out China’s first production-ready, pre-assembled robotaxi developed entirely in-house. It is built on the XPENG GX platform and engineered to L4 autonomous driving standards.
It is a milestone. For the first time in China, a single carmaker has designed, engineered, and manufactured an autonomous taxi entirely in-house.
“The rollout today represents a pivotal step forward in XPENG’s journey from physical AI research and development to large-scale deployment,” the company stated.
As a long-standing rival to Tesla in China, XPENG has developed intelligent EVs specifically designed to challenge the Model 3 and Model Y.
This latest Robotaxi rollout intensifies the direct competition between the two companies as they race for supremacy in both autonomous driving and physical AI.
Advanced computing & tech
Running a massive AI model in a moving vehicle requires substantial power. To solve this, XPENG overcame the limitations of silicon suppliers like Nvidia and built its own.
The Robotaxi is packed with four proprietary Turing AI chips. Together, these chips deliver 3,000 TOPS (trillions of operations per second) of on-chip computing power. It allows the vehicle to instantly analyze its surroundings and execute driving decisions without lagging.
Ditching conventional hardware and data dependencies, XPENG’s Robotaxi operates entirely without LiDAR or high-definition maps.
It utilizes a pure vision solution powered by the VLA 2.0 end-to-end large model. Interestingly, the model compresses system response latency to an ultra-fast sub-80 milliseconds by eliminating the translation step found in legacy “Vision-Language-Action” architectures.
Unlike older, slower self-driving systems that have to translate what they see into computer code before reacting, XPENG’s AI skips the middle step entirely. It simply sees the road and drives instantly.
This streamlined, mapless approach improves the vehicle’s urban generalization capabilities for the self-driving system to adapt to cross-city and even cross-border deployments.
Aimed at delivering a premium, safe, and highly intelligent travel experience, the mass-produced Robotaxi features a luxurious passenger cabin optimized for ultimate comfort. The interior features upscale amenities, including privacy glass, zero-gravity comfort seats, and dedicated rear entertainment screens.
Throughout the journey, passengers can enjoy multimedia content and personalize their environment — such as adjusting temperature or lighting — simply by interacting with the vehicle’s built-in AI voice assistant.
Pilot services
XPENG rapidly accelerated its commercialization timeline early this year by hitting two major milestones in quick succession.
After securing an intelligent connected vehicle permit in January to begin routine L4 public road testing in Guangzhou, the company followed up in March by establishing a dedicated Robotaxi business unit.
This new division centralizes everything from product definition and R&D testing to final operations, streamlining the path from public trials to market deployment.
To evaluate technical performance, market acceptance, and commercial viability, XPENG will debut its pilot Robotaxi services in the latter half of this year.
The company aims to achieve a rapid transition to driverless commercialization, targeting fully autonomous operations without on-site safety officers by early 2027.
“As a full-stack automaker with in-house capabilities spanning software, chips, and complete vehicles, XPENG is positioned to move directly to scaled delivery upon completion of technical validation, thereby shortening the cycle from R&D to commercial operations,” the company stated.
Serving as a flagship pillar of XPENG’s physical AI ecosystem, the mass-produced Robotaxi shares its underlying VLA 2.0 large model with the company’s humanoid robot, IRON, and its flying car.