Browse
Log in / Register
Share
July 20, 2024—Li Xiang, the founder of Chinese electric vehicle maker Li Auto, is aggressively steering the company toward an artificial intelligence future, yet this vision is met with palpable internal confusion and market skepticism. As competitors like Tesla and XPeng accelerate their AI integrations, Li Auto's transition highlights the challenges of shifting from a product-centric to a technology-driven identity.
Internally, Li Xiang's AI advocacy has sparked more questions than clarity. Earlier this year, an unannounced all-hands meeting left employees baffled, with many taking to internal forums to express uncertainty about the company's new direction. This reaction stems from Li Auto's historical focus on short-term, pragmatic goals—such as the successful "refrigerator, TV, and sofa" features that boosted sales—now clashing with long-term AI narratives. Industry observers note that Li Xiang, once hailed as a "hit-making machine" product manager, must evolve into a trusted CTO figure, a transformation that requires convincing his own team first.
The Livis Day event on June 15 was Li Auto's boldest attempt to showcase its AI capabilities. Li Xiang declared that the next decade would focus on "breathing life" into vehicles and homes, moving beyond hardware to embodied AI. The launch featured the in-house Mach M100 chip, boasting 1,280 TOPS of computing power, the Mach Mind large model, and the Mach VLA autonomous driving system. However, the presentation's technical density did little to immediately sway investors, as Li Auto's stock dipped 1.5% post-event and fell another 3.75% the following day, signaling market caution.
Li Auto defines its embodied AI vehicle as a "four-in-one" entity: an EV, a professional chauffeur, an AI computer, and a life assistant. This repackaging of existing concepts—autonomous driving, computing bases, and cabin AI—into a proprietary term aims to differentiate Li Auto in a crowded market. Unlike XPeng's flashy humanoid robot demonstrations, Li Auto's approach is more abstract, emphasizing efficiency and safety over spectacle. Critics argue this is "old wine in new bottles," but Li Xiang insists it represents true intelligence beyond mere function-driven features.
At the core of Li Auto's strategy is the Mach M100 chip, a 5nm automotive-grade processor designed for dynamic data flow. With single-chip computing power of 1,280 TOPS and over 82% operational efficiency, it surpasses Nvidia's Orin-X (254 TOPS) in density. The dual-chip setup in the new L9 Livis model delivers 2,560 TOPS, supporting Li Auto's claim of full-stack in-house development from chips to algorithms. Yet, the real test lies in data scalability: Tesla's FSD V14 benefits from millions of vehicles feeding real-world data, while Huawei's ADS leverages partner networks. Li Auto's assisted driving mileage lags significantly, prompting a reliance on computing power to compensate for data gaps through synthetic simulation training.
Exclusive analysis reveals that Li Auto's AI push coincides with rising talent competition in China's tech sector. According to a recent Bloomberg report, EV firms are poaching AI experts with salaries up to 50% above market rates, a pressure Li Auto addresses through events like Livis Day to retain core algorithm teams. Additionally, Li Xiang's open-sourcing of Xinghuan OS and chip advancements signal a defensive move against short-selling and investor doubts, positioning the company for future capital raises in a volatile market.
Market reception remains mixed. While some early adopters, like an L8 and i6 owner, praise Li Auto's smart driving as "leagues ahead," broader investor sentiment reflects uncertainty about AI commercialization timelines. Li Auto's traditional strength in quick returns—evident in its past product successes—now faces the slower pace of tech innovation. The embodied AI concept, though strategically sound, requires tangible user benefits to gain traction, something Li Auto plans to demonstrate through quarterly OTA updates, with promises to match Tesla's FSD V14 by year-end.
Looking ahead, Li Auto's success hinges on translating AI rhetoric into measurable outcomes. The company must bridge the internal chasm by aligning employee incentives with long-term tech goals, while externally, it needs to prove that embodied AI enhances safety and efficiency beyond competitors' offerings. As the EV landscape evolves, Li Xiang's pivot may redefine Li Auto's identity, but only if it can turn skepticism into sustained innovation in an AI-driven era.
