- China transforms agricultural land into data centers to compete with the American domination of AI
- Wuhu’s $ 37 billion project highlights Beijing’s urgency in artificial intelligence
- The export restrictions let China be based strongly on less powerful local fleas
China’s ambitions in artificial intelligence have acquired new visibility thanks to its plan to develop a national alternative to the Stargate massive project prosecuted in the United States by Openai and Oracle.
While the American initiative is expected to support up to two million IA chips, Beijing increases its own version anchored by a $ 37 billion project in Wuhu.
Although much smaller than the price of $ 500 billion linked to Stargate, the Chinese project is designed to consolidate the existing IT capacity in a more centralized network.
The Wuhu project and its scale
The site selected for this project is in Wuhu, in eastern China, and it covers old rice fields along an island of 760 acres in the Yangtze river basin.
This land, formerly devoted to food production, is converted into “data island” for four of the country’s largest technological operators: Huawei, China Mobile, China Telecom and China Unicom.
By situating the new “mega-cluster” of data centers near major cities such as Shanghai, Hangzhou and Nanjing, planners hope to provide faster inference services to dense urban populations.
From 2022, China encouraged the construction of servers in the interior provinces with cheap power supplies.
However, these sites have often remained inactive, because local governments have reassured the capacity of areas where demand was higher.
The new plan tries to solve this problem by connecting urban and remote data centers via UB-MOSH technology from Huawei.
This technology can provide redundancy while allowing the sale of unused calculation.
The subsidies of the Wuhu project, which would cover up to 30% of IA Chip supply costs, also reflect Beijing’s urgency to make new clusters operational.
China is currently holding around 15% of the world’s computing power power, much less than the 75% of the United States.
Export restrictions have blocked access to the advanced NVIDIA GPUs, leaving national suppliers incapable of fully matching foreign performance.
This gap has created incentives for the smuggling of equipment, although those responsible seem determined to develop self -sufficient AI batteries to reduce dependence on sources abroad.
The long -term objective is that this infrastructure will allow companies and individuals to deploy more sophisticated AI tools.
The question of whether local fleas can support this ambition remain uncertain compared to Western options supplying the main data centers abroad.
The conversion of agricultural land into a server space raises questions about sustainability, resource allocation and demand for energy.
Supporters consider projects essential to reduce the technological fracture, while skeptics highlight the costs of embezzlement of agricultural land and the uncertainty of relying on less powerful local chips.
Via Toms equipment
