The 2026 Beijing Auto Show has highlighted a decisive strategic shift among Chinese electric vehicle manufacturers: the aggressive move toward developing in-house AI chips. As reported by the South China Morning Post, companies like NIO, Xpeng, and Hesai are no longer content to rely solely on foreign suppliers like Nvidia and Qualcomm for the silicon that powers their intelligent driving systems. Instead, they are betting the house on proprietary chip designs to make their cars smarter, more autonomous, and less vulnerable to geopolitical supply chain disruptions.
This trend was front and center at the auto show, where the integration of advanced AI capabilities into vehicle architectures was the dominant theme. The push for in-house silicon is driven by the need for tighter integration between hardware and software, enabling automakers to optimize performance for their specific autonomous-driving algorithms. By controlling chip design, these companies aim to achieve higher efficiency, lower power consumption, and faster iteration cycles than with off-the-shelf solutions.
NIO’s 5nm Shenji NX9031 and Xpeng’s Turing Chip
NIO made waves just prior to the show by launching the 2026 Onvo L90, the first production vehicle to feature its proprietary 5nm Shenji NX9031 chip. The company also showcased its flagship ES9 SUV equipped with the same silicon. The Shenji NX9031 represents a significant milestone, delivering computing power equivalent to four Nvidia Orin X chips and is designed specifically to handle the massive data-processing requirements of NIO’s advanced driver-assistance systems (ADAS). NIO’s CEO has previously called for industry-wide standardization of AI chips to unlock billions in savings, but the company is clearly pushing ahead with its own proprietary advantage.
Xpeng is following a similar trajectory, heavily promoting its Turing AI chip architecture. Xpeng’s approach emphasizes “physical AI,” framing the vehicle as an embodied AI agent rather than just a mode of transportation. The Turing chip is designed to support Xpeng’s VLA 2.0 (Vision-Language-Action) autonomous driving engine, enabling the car to process complex visual and linguistic inputs in real-time. This tight coupling of proprietary silicon and advanced AI models is becoming the new baseline for premium Chinese EVs.
The Supply Chain Security Imperative
Beyond performance optimization, the shift toward in-house chips is fundamentally about supply chain security. With the US government continuously tightening export controls on advanced semiconductors, Chinese automakers recognize the existential risk of relying on foreign silicon for their most critical vehicle systems. Developing domestic chip design capabilities, even if the manufacturing still relies on foundries like TSMC, provides a crucial buffer against future geopolitical shocks.
Lidar manufacturer Hesai also demonstrated this trend, showcasing its next-generation sensors that increasingly rely on localized processing power. As the architecture of the software-defined vehicle evolves, the line between automaker and semiconductor company is blurring. The 2026 Beijing Auto Show has made it clear: in the race for autonomous driving supremacy, Chinese EV makers believe that controlling the silicon is just as important as controlling the software.