There is little doubt that the artificial intelligence race today has two major protagonists: the United States and China. They are not the only countries that are moving forward, but they are the two that are setting the pace, each with their own tools and with a very different idea of how to sustain progress. As AI begins to become economic infrastructure, the question changes. It’s no longer just a question of who has the best models, but of who can build the material base to power them, deploy them, and take them everywhere.
First idea. Large-scale state financing. According to Bloomberg, China is preparing a plan to allocate around 2 trillion yuan, about $295 billion, over the next five years to build AI data centers across the country. The information points to a deployment promoted by Beijing to strengthen its national artificial intelligence sector. We are not yet talking about a closed plan: the media points out that the project is still in an early phase of discussion and that the details may change.
One network, not just more data centers. The key to the plan would not only be in building new facilities, but in connecting them under a national architecture. Bloomberg talks about a network of interconnected computing hubs that would allow resources currently dispersed between regions to be pooled and give companies and organizations broader access to high-performance capacity. The general objective would be for these facilities, now fragmented, to function as a more cohesive system by 2028.
The State as architect. At the center of the design are organizations such as the National Development and Reform Commission, one of China’s great economic planning arms. On the other hand, state companies such as China Mobile and China Telecom would assume a good part of the operation of these centers and the connectivity between them. It is an important detail because it helps understand Beijing’s approach: it seeks to position itself as a coordinator, according to the information released by the American media.
Second idea. The other big leg of the plan is who would supply the technology. Bloomberg notes that the idea is to turn to local suppliers, including Huawei, for at least 80% of the hardware and software, including AI chips. That threshold does not amount to an explicit ban on NVIDIA or AMD, but it would leave them with very little room to participate in the rollout. It is precisely there where investment also becomes a tool to reduce foreign technological dependence.
It is not an isolated movement. The direction fits with steps that Beijing had already been taking to reduce dependence on foreign chips in private and public infrastructure. Without going any further, the market share of the firm led by Jensen Huang has plummeted in recent months, and there is little reason to think that it could rise again soon.
The background signal. It should be noted that the plan advanced by Bloomberg is not officially confirmed, but it shows where Beijing wants to move if it finally goes ahead. China would not only be preparing a huge investment in data centers: it would be trying to make this deployment work as a national network, largely powered by local technology.
Images | with Nano Banana
In | We already know how much water Amazon consumes in its data centers. We have good and bad news
