Chinese EV manufacturer Xpeng (NYSE: XPEV) announced plans to introduce three robotaxi models in 2026, marking a major step in its autonomous mobility strategy. Trial operations for the service will begin the same year.
Each robotaxi will utilize four of Xpeng's proprietary Turing AI chips, providing a combined computing capacity of up to 3,000 TOPS. These vehicles will also feature Xpeng’s second-generation VLA (Vision-Language-Action) model, enabling adaptive performance in diverse and complex traffic environments worldwide.
The new models are designed for mass production without requiring physical modifications or reliance on high-definition maps. This approach significantly reduces costs and supports deployment beyond limited zones, including intercity and international operations.
The robotaxis will include external communication features that let pedestrians see the car’s immediate intentions—such as turning or searching for parking—displayed through the windshield interface.
Xpeng plans to open its robotaxi Software Development Kit (SDK) for partners worldwide, aiming to establish a collaborative ecosystem to accelerate the adoption of autonomous ride-hailing solutions.
He Xiaopeng, chairman and CEO of Xpeng: “These robotaxis integrate our latest Turing AI and VLA systems to make urban mobility safer, more efficient, and globally scalable.”
Author’s summary: Xpeng will release three robotaxi models in 2026 equipped with advanced in-house AI chips and VLA technology, aiming for global scalability and low-cost autonomous mobility.