Li Auto, a prominent player in the electric vehicle (EV) industry, has made significant strides in the development of its in-house autonomous driving chip. This breakthrough comes after other domestic EV manufacturers like Nio Inc and Xpeng have successfully integrated their own chips into mass-produced vehicles.
The M100 autonomous driving chip by Li Auto reached a crucial milestone in the first quarter of this year with the completion of prototype production. This achievement was followed by thorough functional and performance testing, which included rigorous stress tests conducted by the company’s research and development engineers.
Currently, the M100 chip is undergoing road testing on prototype vehicles after being installed in small batches. Notably, the chip has shown impressive computing power capabilities, particularly in tasks involving large language models (LLMs) and traditional vision tasks utilizing convolutional neural networks (CNNs).
According to a report by LatePost, a single M100 chip can deliver computing power equivalent to two Nvidia Thor-U chips for LLM tasks and three Thor-U chips for CNN tasks. This showcases the efficiency and performance capabilities of Li Auto’s autonomous driving chip.
Looking ahead, Li Auto plans to commence mass production and integrate the M100 chip into its vehicles next year. Until then, the company will continue to collaborate with existing partners such as Nvidia and Horizon Robotics for smart driving chip solutions.
The trend of EV manufacturers developing in-house chips aligns with the need to tailor hardware to proprietary algorithms and product release schedules. Nio, for instance, introduced its Shenji NX9031 smart driving chip in late 2023 and has since incorporated it into various models, including the ET9 and ES8.
Similarly, Xpeng launched its Turing AI chip in the G7 SUV and P7 electric sedan, showcasing the company’s commitment to in-house chip development. Li Auto’s substantial investment in its chip project encompasses not only hardware components like NPUs and SoCs but also software development and adaptation.
Under the leadership of CTO Xie Yan, who brings a wealth of experience from companies like Intel, Alibaba, and Huawei, Li Auto aims to leverage software scheduling capabilities to maximize chip hardware computing power. This integrated hardware-software R&D strategy positions the company to outperform competitors in terms of performance and efficiency.
As the EV industry continues to evolve, advancements in autonomous driving technology, particularly in the realm of in-house chip development, will play a crucial role in shaping the future of electric vehicles. Li Auto’s progress with the M100 chip underscores its commitment to innovation and technological excellence in the pursuit of delivering cutting-edge EV solutions to consumers.
