Hurun Research Institute’s 2025 China Artificial Intelligence Top 50 list, released January 19, 2026, ranks AI chip maker Cambricon first with an estimated value of 630 billion yuan, up 165% year-on-year. GPU unicorns Moore Threads and Muxi rank second and third, underscoring how compute hardware dominates the upper tier of China’s AI ecosystem.
This article aggregates reporting from 1 news source. The TL;DR is AI-generated from original reporting. Race to AGI's analysis provides editorial context on implications for AGI development.
Hurun’s new ranking is a snapshot of where value is accruing in China’s AI stack: overwhelmingly in compute. Seven of the top ten companies are chip or hardware players, led by Cambricon, Moore Threads and Muxi, with valuations that have grown triple digits in a year. That tells you two things: first, Chinese capital markets believe domestic AI chips can partially substitute for restricted Nvidia parts; second, investors see a long runway of demand from both model training and an emerging wave of edge and industrial deployments. ([21jingji.com](https://www.21jingji.com/article/20260119/herald/126fddc51aec25ba21528b9cd3ae747d.html))
For the race to AGI, this hardware-heavy leaderboard underscores how different China’s ecosystem looks compared to the US, where the highest private valuations sit with model labs and cloud platforms. If Chinese GPU and accelerator vendors can deliver competitive performance and scale manufacturing, they could underpin a parallel AI universe with its own frontier models, tuned first for domestic languages, regulations and industrial needs. At the same time, the presence of model companies like iFlytek, Baichuan, Moonshot AI and autonomous driving players such as Pony.ai and Momenta in the top 50 shows that China’s software and embodied AI bets are tightly coupled to this compute base.
The strategic implication: even if export controls limit access to the very latest Nvidia parts, a deep bench of domestic chip startups—backed by rising paper valuations—gives China multiple shots at building the infrastructure necessary for homegrown AGI-scale experiments.

