China is accelerating its artificial intelligence ambitions with a massive data centre project in Wuhu, eastern China, intended to rival the United States’ Project Stargate. The initiative, valued at US$37 billion, reflects Beijing’s determination to consolidate computing power and develop alternatives to Western technology despite facing significant supply chain hurdles.
A domestic answer to Project Stargate
The American Project Stargate, led by OpenAI and Oracle, has been billed as a computing initiative capable of supporting up to two million AI chips, with an estimated cost of US$500 billion. While China’s project is far smaller in scale, its goal is similar: to centralise AI computing resources into a nationwide infrastructure that can support rapid deployment of artificial intelligence.
The Wuhu site, chosen for its proximity to Shanghai, Hangzhou and Nanjing, is being developed on a 760-acre island in the Yangtze River basin. Once used for rice farming, the land is now being transformed into what officials call a “data island”. Four of China’s largest technology companies—Huawei, China Mobile, China Telecom and China Unicom—are leading the project, which will operate as a mega-cluster of data centres to deliver faster AI services to densely populated urban areas.
China had previously encouraged the building of server farms in interior provinces where energy was cheaper, starting in 2022. However, many of these sites remained underused because local governments diverted resources to regions with greater demand. The new approach seeks to solve this mismatch by connecting both urban and remote facilities through Huawei’s UB-Mesh technology, which provides redundancy and enables unused computing capacity to be resold.
Subsidies and challenges in hardware
The Wuhu project benefits from substantial government support, with subsidies reportedly covering up to 30 per cent of AI chip procurement costs. These incentives highlight Beijing’s urgency to ensure the new infrastructure is operational as soon as possible.
Currently, China controls approximately 15% of the global AI computing capacity, lagging behind the United States, which is estimated to hold around 75%. Export restrictions imposed by Washington have cut off access to advanced Nvidia graphics processors, creating significant limitations for Chinese firms. This has led to concerns about hardware smuggling and a renewed push for domestic chip innovation. Officials are intent on building a self-sufficient technology stack that reduces reliance on foreign suppliers, though it remains unclear whether local chips can yet match Western standards.
The long-term goal is to enable both companies and individuals across China to deploy more advanced AI tools. Achieving this vision will depend heavily on the success of domestic chipmakers in narrowing the performance gap with international competitors.
Concerns over farmland and sustainability
The transformation of rice fields into server space has raised questions about sustainability and land use priorities. Critics argue that diverting agricultural land for technology development could threaten food security and place additional strain on energy resources. The new data centres will also consume vast amounts of electricity, raising concerns about environmental impact.
Supporters of the project argue that such initiatives are crucial for China to narrow the technological gap with the United States and maintain its position in the AI race. For them, the strategic importance of building homegrown computing power outweighs the risks associated with land conversion and energy demand.
As the Wuhu development progresses, the balance between technological advancement and sustainable resource utilisation is likely to remain a central debate in China’s pursuit of AI dominance.
