Bloomberg broke the details yesterday: China is preparing an AI infrastructure program that dwarfs anything we’ve seen before. 2 trillion yuan — roughly $295 billion — earmarked for nationwide AI data centers over the next five years.
The Scale
$295 billion. For context, that’s more than Finland’s entire GDP. State-owned enterprises like China Mobile and China Telecom will operate the data centers. If power grid upgrades are included, the total could exceed 5 trillion yuan.
The goal is ambitious: by 2028, China wants a unified national computing environment. Not a patchwork of regional projects, but a coordinated system.
Huawei, Not Nvidia
Here’s the geopolitically explosive part: Huawei will supply over 80% of the technology. The US export restrictions against China have produced exactly what many feared — not less AI development, but an accelerated decoupling from Western tech.
Nvidia and AMD, already constrained by export controls, are being pushed even further out of the world’s second-largest AI market. For Nvidia especially, this stings — China was one of their most important markets before the sanctions.
What This Means
This isn’t wishful thinking. Bloomberg reports that China’s State Council has already approved the plans. The funding is in place, contracts are being assigned.
For the global AI landscape, the implications are significant. Two separate technology ecosystems are becoming reality — one built around Nvidia and US cloud providers, one around Huawei and Chinese state enterprises. The question isn’t whether decoupling happens, but how fast.
My take: The export restrictions were supposed to slow China’s AI development. Instead, they’re accelerating the build-out of a parallel infrastructure. It’s a pattern we see with sanctions over and over — short-term impact, long-term competition. Whether Huawei’s chips can keep up with Nvidia’s H100/B200 generation remains an open question. But $295 billion buys a lot of development time.
Sources: Bloomberg, TechStartups, Dataconomy