China’s $295 Billion AI Plan — and What It Means for the Global AI Race

On June 9, Bloomberg reported that China is preparing to spend 2 trillion yuan — approximately $295 billion — on AI infrastructure over the next five years. The plan involves a nationwide network of interconnected data centres and computing hubs designed to support AI model training and deployment at scale.

It is one of the largest technology infrastructure programmes ever announced by any government. And it changes the shape of the global AI race in ways that go well beyond the headline number.

What the Plan Actually Involves

The blueprint is being drafted by China’s National Development and Reform Commission alongside other key government agencies. The goal is to create a unified computing environment by 2028, which would support AI resources now distributed across regions and organisations.

State-owned telecommunications giants, including China Mobile and China Telecom, are expected to operate most of the planned facilities. Officials reportedly want at least 80% of the hardware used across the network — including AI processors — to come from domestic suppliers. That means Huawei, Alibaba’s chip division, Biren Technology, and Moore Threads. Not Nvidia, not AMD, not any American company.

The plan also extends beyond data centres. Reports indicate that

communications infrastructure and power systems may be integrated into the broader buildout. If electricity grid upgrades are included, total investment associated with the project could reportedly exceed 5 trillion yuan. At that scale, this is not just an AI project. It is a national infrastructure transformation.

Why China Is Doing This Now

The timing is not a coincidence. US export controls have cut China off from Nvidia’s most advanced AI chips — the H100 and H200 series that power most of the world’s frontier AI development. China cannot buy its way to compute parity with the West the way it once could. So it is doing the next best thing: building its own supply chain and deploying it at a scale that brute-force compensates for the chip gap.

Recent approvals granted in May 2026 to nine categories of domestically developed AI chips from companies including Huawei, Alibaba, Biren Technology, and Moore Threads have opened the door for broader deployment across government and security-sensitive sectors. Those approvals give the plan its hardware foundation. The $295 billion gives it financial momentum.

There’s also a domestic economic dimension. China is facing significant headwinds — mounting government debt, a property sector that hasn’t fully recovered, and slowing export growth. A massive state-backed infrastructure programme serves both as an AI strategy and as an economic stimulus. The two goals reinforce each other.

How It Compares to US Spending

The numbers look large. But context matters. American technology companies are projected to spend more than $700 billion this year alone on AI infrastructure. That’s in a single year — compared to China’s $295 billion over five years.

The difference is structural. In the US, AI infrastructure spending is driven by private companies — Meta, Microsoft, Amazon, Google — competing with each other and spending aggressively to win. In China, the approach is state-coordinated, more measured, and focused on building shared national infrastructure rather than competing commercial clouds.

Neither approach is obviously better. The US private-sector model moves faster and tolerates more experimentation. The Chinese state-backed model is better at standardisation and long-term resource allocation. The question is which approach produces better AI capabilities over a five-to-ten year horizon — and that question is genuinely open.

The Chip Problem China Still Has to Solve

The most pointed criticism of the plan is that capital alone can’t solve China’s chip constraints. Huawei’s most advanced AI chip — the Ascend 910C — is competitive with Nvidia chips from a few generations ago, but still trails the current state of the art. Building more data centres filled with slower chips doesn’t close the gap. It scales it.

Analysts at Digg noted that China’s bottleneck isn’t data centre construction — it’s chip performance. Money can build buildings and lay cables. It cannot easily accelerate the semiconductor physics that determine how fast AI models train. TSMC, which manufactures the world’s most advanced chips, is off-limits to Chinese companies under US export controls. That’s a ceiling that $295 billion in data centre spending doesn’t raise.

That said, China has surprised Western analysts before. Huawei’s 7nm chip, revealed in 2023, was considered impossible given export restrictions. China’s AI research output has grown dramatically. Dismissing this plan as simply throwing money at an unsolvable problem may be exactly the kind of complacency that leads to being caught off guard.

What It Means for the Global AI Race

The AI race is no longer primarily a competition between companies. It is a competition between ecosystems — and increasingly, between geopolitical blocs.

The US bloc (including the UK, EU, Japan, South Korea, and Taiwan) controls the most advanced chip manufacturing, the leading AI models, and the dominant cloud platforms. The China bloc is building its own version of all three — slower in some respects, but at a scale and coordination level that no private company can match.

The risk for Western technology companies is not that China overtakes them tomorrow. It’s that a fully self-sufficient Chinese AI ecosystem — built on domestic chips, domestic cloud, and domestic models — creates a permanent split in the global technology stack. Two internets. Two AI supply chains. Two sets of standards. That bifurcation is already underway. The $295 billion plan accelerates it significantly.

The AI race is no longer about who has the best chatbot. It’s about who controls the compute, the power, and the supply chain behind the next generation of AI systems.

What to Watch

This plan is still in draft form. No fixed allocations or milestones have been confirmed yet. The real test will come in the next 12 to 18 months, as project contracts are awarded and construction begins. Watch which domestic chip companies receive the largest orders — that will reveal how serious Beijing is about the 80% domestic content requirement.

Also watch the power grid dimension. If electricity infrastructure is formally included in the plan’s scope, the $295 billion figure will be revised significantly upward, and the project will take on a different strategic character — less AI initiative, more national infrastructure transformation.

Either way, the era of AI as a purely corporate competition is over. Governments are now major players in the infrastructure race. And China just placed a very large bet. For more on how the AI arms race is reshaping global tech, see our breakdown of the emerging technologies transforming the future — and what they mean for businesses and individuals alike. For a closer look at how AI is changing the tools we use every day, see our guide to the best free AI tools available right now.

Read more tech related articles here.