China is planning a $295 bn AI data‑centre build that blocks Nvidia chips.
Beijing is drafting a five‑year blueprint to spend roughly 2 trillion yuan on AI‑specific data centres. The plan calls for dozens of new facilities across the country, each designed to run models on domestically produced hardware. Official documents explicitly mention excluding foreign GPU suppliers, targeting Nvidia in particular.
The move could reshape the global AI supply chain by creating a parallel ecosystem that relies on Chinese chips. It also raises the cost for multinational hardware firms trying to stay in the Chinese market, and may force AI developers to maintain two separate model stacks.
If the plan survives political wrangling, it will be another costly attempt to nationalise AI talent, but history suggests technology rarely stays locked behind borders.