Shanghai-based MetaX (known in Chinese as Muxi) was founded in 2020 by three former AMD veterans and has already secured R&D teams across Beijing, Nanjing, Chengdu, Hangzhou, Shenzhen, and Wuhan. The company recently passed a regulatory hearing for listing on the Shanghai Star Market, a sign of how China is intensifying its push into domestic AI-chip design, says the South China Morning Post.
MetaX focuses on designing chips rather than owning fabs (i.e., it is fabless). Its product lineup includes:
- “N-series” GPUs for AI inference
- “C-series” general-purpose processors
- “G-series” graphics rendering units
A key piece: MetaX’s platform, called MXMACA, is built to be compatible with mainstream GPU ecosystems like Nvidia’s CUDA, a move aimed at easing adoption by developers familiar with the dominant stack.
In July, MetaX launched the C600 general-purpose processor. This chip is built with HBM3e (high-bandwidth memory) and supports FP8 precision (used widely in AI training/inference) and reportedly sports 144 GB of memory. MetaX describes it as “fully domestically produced” and plans mass production by year-end.
Why this matters: Nvidia retains a large lead in global AI GPUs, but China is under export restrictions and wants a home-grown alternative. MetaX embodies that shift, offering a design ecosystem, domestic supply chain posture, and government-backed ambition. That said, hurdles remain: commercial-scale production, benchmark performance parity with Nvidia, software ecosystem maturity, and global supply-chain integration, all still pose serious tests.
MetaX is more than just a chip startup. It’s part of a national tech-strategy push to challenge established GPU incumbents and build a domestic engine for AI hardware.