The United States has its executive orders. The EU has its AI Act. But on March 5, 2026, China unveiled something far more sweeping: a 141-page national blueprint that essentially bets the country’s economic future on artificial intelligence.

China’s 15th Five-Year Plan, released at the opening of the National People’s Congress in Beijing, mentions AI more than 50 times and includes what officials are calling an “AI+ action plan” — a comprehensive strategy to embed artificial intelligence into virtually every sector of the world’s second-largest economy. From manufacturing floors to hospital wards, from logistics networks to classrooms, Beijing is telling the world: AI isn’t a side project. It’s the project.

And given what Chinese AI companies like DeepSeek have accomplished in the past year, dismissing this as aspirational paperwork would be a serious mistake.

What’s Actually in the Plan

The “AI+ action plan” is the centerpiece. It calls for deploying AI agents across industries with minimal human guidance. It envisions robots filling jobs in sectors suffering from labor shortages — not as a futuristic concept, but as a near-term policy experiment. The government is committing to “hyper-scale” computing clusters powered by cheap, plentiful electricity to train the next generation of models.

Premier Li Qiang’s government work report highlighted technology — which Beijing calls “new quality productive forces” — in its opening paragraphs, far more prominently than in 2025. The message is unmistakable: this is priority number one.

Three Forces Driving China’s AI Bet

The demographic time bomb. China’s population has declined for four consecutive years. Its workforce is aging rapidly, and unlike the US or Europe, China can’t import workers at scale. The solution? Replace them with machines. The plan explicitly treats AI and robotics as a demographic survival strategy.

The DeepSeek effect. Chinese AI developers have shocked the industry. DeepSeek’s models have gone toe-to-toe with Western competitors at a fraction of the cost. This gave Beijing genuine confidence that its ecosystem can compete — and potentially lead — globally. The state planning body now boldly claims China “leads the world in research and development and application in fields such as AI, biomedicine, robotics and quantum technology.”

The tech war with Washington. US export controls on advanced chips have been a source of immense frustration. But rather than backing down, China is doubling down on self-reliance. The plan emphasizes independent chip R&D and domestic innovation, treating American restrictions as motivation. As one expert noted, “If the last Five-Year Plan’s innovation policy was largely defensive, this one is much more proactive.”

Beyond AI: Quantum, 6G, and Humanoid Robots

The plan’s ambitions extend into territory that sounds like science fiction.

China is committing to scalable quantum computers and an integrated space-earth quantum communication network — essentially a hack-proof system linking satellites and ground stations. It’s pledging breakthroughs in nuclear fusion, reusable heavy-load rockets, and feasibility studies for a lunar research station.

On robotics, the plan pours resources into “embodied AI” powering humanoid robots. Combined with 6G networking and brain-machine interfaces, Beijing is sketching an economy where physical and digital intelligence merge at a level no other country has attempted to plan centrally.

The Open Source Wildcard

A detail that flew under the radar: for the first time, China’s national plan explicitly supports open-source AI communities.

“Open source wasn’t mentioned before in previous reports and is also a key difference between the Chinese AI and American AI approaches,” noted Tilly Zhang at Gavekal Dragonomics.

While US development is increasingly dominated by closed models from OpenAI and Anthropic, China wants a collaborative, open ecosystem. Given that DeepSeek and others have already embraced open-weight releases, this could accelerate a global shift in how models are developed and shared.

It’s also strategic. Open-source AI gives China influence over global AI infrastructure without requiring direct market access in Western countries. If developers worldwide build on Chinese open-source foundations, Beijing gains soft power in the AI stack even where its companies can’t directly compete.

What This Means for the US-China AI Race

This plan drops at a moment when the US AI strategy looks fragmented. Washington is juggling Pentagon AI contracts that have become political lightning rods, energy infrastructure that can’t keep pace with demand, and a public backlash against AI companies showing no signs of slowing.

China, by contrast, is presenting a unified national vision backed by the full weight of state planning. That doesn’t mean it’ll succeed — centrally planned initiatives often stumble in execution, and chip manufacturing still lags behind TSMC. US export controls remain a genuine bottleneck.

But the ambition gap is real. The US has no equivalent of an “AI+ action plan.” No 141-page federal blueprint for integrating AI into every sector. Individual companies are charging ahead, but national coordination is virtually nonexistent.

The question isn’t whether China will overtake the US in AI. It’s whether a state-directed approach to AI deployment can move faster than a market-driven one when the stakes are this high. China is betting yes. And with 1.4 billion people and a government that doesn’t need to win midterm elections, it’s a bet worth taking seriously.

A New Kind of Superpower Competition

What makes this historically significant isn’t just the AI content — it’s what it reveals about great power competition.

The Cold War was about nuclear weapons and ideology. The current US-China competition is about who can build and deploy advanced AI faster and more effectively. Whoever gets AI integration right first — not just impressive models, but embedding them into economic infrastructure at scale — gains a compounding advantage that could define the century.

China’s plan is explicit about this. It’s not chasing AI capability. It’s chasing AI application — the unsexy work of deploying systems in factories, hospitals, ports, and farms. That’s arguably harder than building the models, and it’s where centralized planning could prove a genuine advantage.

Whether you find that exciting or terrifying probably depends on where you sit. But ignoring it isn’t an option.