Li Auto's chip unit Mach Core raises $16.25B RMB 1.63B first round at ~16.6B valuation
创业邦 · wechat · 2026-10-18
Li Auto's chip subsidiary Shanghai Mach Core signed its first equity financing of RMB 1.625B: RMB 1B from Li Auto's system, RMB 625M from outsiders incl. Lanchi Ventures (co-lead), Hangzhou Jintao, Huaye Tiancheng and MiraclePlus. Pre-money valuation is RMB 15B, post-money RMB 16.63B; an additional RMB 500M option could push it to RMB 17.1B and dilute Li Auto to 74%. The deal includes 15% employee and 5% executive incentive pools.
Note that Mach Core, incorporated Aug 2026, has no operations or revenue yet; the already-mass-produced M100 (5nm automotive AI inference chip) remains under Li Auto's original R&D and asset transfer is still ongoing. The chip project started Nov 2022 and reached mass production in 3.5 years with a 200-person team; former SoC lead Qin Dong left for AI chip startup YiZhu.
The article analyzes the commercialization challenge: like Nio's Shenji unit (RMB 2.2B+ first round, RMB 10B post-money), both carmakers keep control while seeking external sales, with robotics as a key target—but external customers must weigh model migration and toolchain costs against incumbents like Nvidia, Horizon Robotics, Black Sesame and Axera; robotics demand ultimately needs production orders to validate.
Related event: Li Auto's chip unit Mach Heart closes 1.63B yuan first funding round(2 posts)→
More from Companies & People
- Ex-Tesla engineer: a month on the factory floor taught more than any elite school — yunta_tsai · 2026-10-18
- Walmart's billion-dollar warehouse robots can't handle real-world work, WSJ reports — ZehanWang · 2026-10-18
- Epic CEO Tim Sweeney urges devs to openly embrace AI: most already use and love it — lucasmeijer · 2026-10-18
- Bret Taylor resurfaces Mike Cassidy's classic talk on speed as THE startup strategy — soleio · 2026-10-18
- Stack Overflow salary median rose 5% in a year, but the same developer got 1% — dima806_dima · 2026-10-18
- AI Book Club hosts Nathan Lambert on RLHF and why labs are scaling RL — sophiamyang · 2026-10-18