Li Auto M100 Chip / 马赫 M100

Overview

Li Auto M100 Chip / 马赫 M100 is the inference chip named by 李想 in 李想×罗永浩！李想的理想：通过 AI 技术，让普通人也过上富豪的生活 as a key component of 理想汽车’s AI and embodied-intelligence strategy.

Current Profile

The source positions M100 as both a vehicle compute upgrade for L9 LEVIUS and a longer-term platform bet. Li argues that after using third-party assisted-driving chips, Li Auto saw a need for its own inference architecture, toolchain, stability, and cost profile.

Key Characteristics

• Automotive inference chip presented as part of Li Auto’s self-developed physical-AI stack.
• Connected to L9 LEVIUS compute, model size, frame rate, vehicle lifecycle upgrades, and future L3/L4 readiness.
• Framed as a dynamic-data-flow direction rather than a generic GPU-style endpoint.
• Treated by Li as a long-term moat even at the cost of near-term profit.
• Source-scoped product whose performance and adoption need independent proof.

Evidence

• Vehicle integration: 李想×罗永浩！李想的理想：通过 AI 技术，让普通人也过上富豪的生活 says L9 LEVIUS uses M100 and reaches 2560 TOPS compute.
• Investment rationale: 李想×罗永浩！李想的理想：通过 AI 技术，让普通人也过上富豪的生活 has Li say the chip effort followed prior experience with Mobileye IQ4, Horizon Journey chips, and Nvidia Orin X.
• Moat thesis: 李想×罗永浩！李想的理想：通过 AI 技术，让普通人也过上富豪的生活 says future embodied intelligence will make inference chips central, while training-chip advantage is already strongly associated with Nvidia.

Qualifications

The source is a founder interview. It does not benchmark M100 against competing automotive inference chips, verify the claimed architecture advantage, or prove that the chip will become a durable moat.

What Changed

• Added M100 as its own entity because it is central to the episode’s hardware and embodied-intelligence strategy.

Relationships

• 理想汽车 - developer and user of the chip.
• 理想 L9 LEVIUS - vehicle platform where the chip is named.
• Inference Chip as Embodied Moat - strategic thesis attached to the chip.
• AI Chip Specialization - broader chip-specialization concept.
• AI Inference Cost Structure - economics context for inference hardware.