Chinese artificial intelligence giant SenseTime has officially unveiled its ambitious Galaxy project. This initiative brings together nearly two dozen leading chip and infrastructure equipment manufacturers from China. The goal is to build five powerful computing clusters, each equipped with 10,000 Chinese AI accelerators.

Scale and Infrastructure

The total capacity of the clusters will be 50,000 specialized chips, making this project one of the largest in China not reliant on American technology. In addition to computing power, Galaxy includes the creation of a token hub for efficient data and model management. SenseTime also plans to support 200 AI startups by providing them with access to this infrastructure.

Notably, the company has not yet disclosed either the construction timeline or the investment amount. This suggests the project is in an early stage of coordination, although its scale is already impressive.

Strategic Context

The launch of Galaxy is a direct response to the tightening of export restrictions by the United States, which block the supply of advanced AI accelerators from NVIDIA and AMD to China. SenseTime, like other Chinese players, is forced to transition to domestic chips, even if they potentially offer lower performance. Creating clusters of 10,000 accelerators each is not only a technological challenge but also a demonstration of China's drive for technological self-sufficiency.

My expert perspective: The Galaxy project is a bold step, but its success directly depends on the real-world performance of Chinese chips. If they cannot compete with Western counterparts in training large models, the clusters risk becoming an expensive but inefficient solution. Nevertheless, for the domestic market and startups, this could provide a powerful boost.