SenseTime’s Galaxy Project Targets AI Chip Scale-Up
Meta description: SenseTime’s Galaxy Project unites nearly 20 partners to scale domestic AI chip infrastructure in China. Here’s what it means for the industry.
China’s AI sector just got a major infrastructure push. SenseTime has unveiled its Galaxy Project, a sweeping initiative that brings together nearly 20 partners to accelerate the country’s shift toward domestically produced AI computing power. For an industry racing to reduce dependence on foreign chips, this is a signal moment.
What the Galaxy Project Actually Involves
A Closed-Loop Ecosystem
At a keynote titled “Intelligent Transformation and Symbiosis,” SenseTime co-founder and Large Device Business Group president Yang Fan outlined a closed loop connecting chip-level technology, ecosystem partnerships, and commercial deployment for domestically produced AI computing power. In plain terms, SenseTime wants to link chipmakers, component suppliers, and infrastructure firms into one coordinated supply chain rather than isolated efforts.
The domestic chip ecosystem backing this project reads like a who’s-who of Chinese semiconductor players. Partners include Cambricon, Muxi, Hygon, Huawei Ascend, Moore Threads, Sunrise, and Biren Technology, alongside infrastructure firms such as Silicon Motion and Zhongke Jiahe.
Scale and Ambition
The numbers behind Galaxy are substantial. SenseTime’s plan covers construction of one “token factory,” five computing clusters described as “10,000-calorie” scale, coordinated work spanning ten technology directions, and support for 200 AI startups. Separately reported figures suggest each cluster could house roughly 10,000 Chinese-made AI accelerators apiece, though SenseTime hasn’t disclosed a construction timeline or budget yet. ForkLog
Why Domestic AI Chip Scale-Up Matters Now
Three Converging Trends
Yang framed the timing around a specific convergence of forces. He pointed to climbing enterprise token demand, industrial AI adoption finally catching up with consumer applications, and domestic chip commercialization reaching a maturity point. Together, these trends suggest China’s homegrown chip ecosystem is transitioning from experimental to production-ready.
Solving the Fragmentation Problem
One of the biggest headaches in China’s chip landscape has been fragmentation — different manufacturers running different architectures, compilers, and libraries, forcing developers to re-adapt models for each new chip. SenseTime’s broader strategy, including prior work adapting its Rixin multimodal models across hardware, aims to abstract this complexity into a unified software layer so switching between chip brands becomes seamless.
Beyond Chips: Space and Research Partnerships
SenseTime isn’t stopping at ground-based infrastructure. The company also signed a space computing agreement with satellite manufacturer Guoxing Aerospace and formed a research partnership with five institutions, including the Shanghai Artificial Intelligence Laboratory, focused on scientific computing. It’s a notable diversification bet for a company that has posted years of steep losses, signaling SenseTime wants to be seen as infrastructure-critical, not just another AI vendor.
Conclusion: A Bet on Coordinated Scale
The Galaxy Project is less about any single chip breakthrough and more about orchestration — stitching together China’s fragmented AI hardware ecosystem into something that functions at scale. Whether it delivers commercially remains to be seen, but it’s a clear marker of where China’s AI infrastructure ambitions are headed. Keep watching this space as more details on timelines and funding emerge.
