China's 15th Five-Year Plan dropped this week and it reads like a tech strategy document. AI appears 52 times — up from 11 in the last plan. The "AI+ action plan" targets integrating AI across 90% of the Chinese economy by 2030. AI-related industries are projected to hit 10 trillion yuan ($1.45 trillion) by the end of the plan period. Beyond AI: quantum computing, humanoid robots, 6G, brain-machine interfaces, nuclear fusion. The growth target dropped to 4.5-5%, down from 5%. But the tech ambitions went up.

1. This Is How You Win a Tech War (Beijing, Huawei, DeepSeek)

The US is bombing Iran. China is rewriting its economy around AI. One of these is a five-year strategy.

The numbers are staggering. 52 mentions of AI. 10 trillion yuan in projected AI industry value. 12.5% of GDP targeted for core digital economy industries. This isn't a policy paper — it's an industrial mobilization. China is doing to AI what it did to solar panels and electric vehicles: pick the sector, flood it with state capital, and scale until nobody can compete on price.

DeepSeek proved the model works even under sanctions. The R1 reasoning model matched OpenAI's GPT-4 at a fraction of the cost, trained for roughly $5.6 million on older Nvidia A800 chips that predate the export controls. It temporarily erased about $1 trillion in US tech market value. The lesson Beijing took: we don't need their chips to build competitive AI. We just need to be more efficient.

The semiconductor push is the insurance policy. Ren Zhengfei is leading 2,000+ companies toward 70% self-sufficiency by 2028. Huawei alone has received over $30 billion from central and Shenzhen governments for chip plants. The government has earmarked up to $70 billion in total chip subsidies. China is building a parallel tech stack that doesn't depend on the US at all.

2. But The Money Just Isn't There (Carnegie Endowment, PitchBook)

China's leadership is confident. Investors are not. The VC ecosystem that funded China's last tech boom is gone.

Carnegie's headline: "It Won't Work." The core argument: China's venture capital ecosystem is dry at the critical moment for AI. VC funding for Chinese AI startups dropped nearly 50% year-over-year in early 2025. In Q2, funding hit $4.7 billion — the lowest in a decade.

The formula that powered China's internet boom is broken. In the mid-2010s, China averaged 238 new VC funds annually. Private investors clustered in coastal cities, catalyzing investment in promising startups. That's how Alibaba, Tencent, and ByteDance scaled. That capital pipeline is gone. VC fund creation has collapsed to less than a quarter of its recent peak.

The 90% target is aspirational, not operational. Merics, the European China think tank, called the plan "self-reliance, diffusion, and a lot of hype." China's all-in embrace of AI is "no guarantee of success." The 90% number is a political signal, not an actionable KPI. And growing scrutiny of cybersecurity will make exporting Chinese AI increasingly difficult.

3. The Export Controls Are Backfiring (ITIF, FPRI)

The US tried to choke China's chip supply. Instead, it accelerated exactly the self-reliance Beijing wanted.

ITIF published it plainly: "Export Controls Helped Huawei and Hurt U.S. Firms." Every new restriction pushes China to build domestic alternatives faster. DeepSeek built competitive models using older chips that weren't covered by controls. The controls didn't stop China — they changed its supply chain.

And now the US wants to expand controls globally. The Trump administration is reportedly considering requiring government approval for ALL AI chip exports worldwide — not just to China. Foreign governments may need to invest in US data centers or provide security guarantees for large orders. If the old controls backfired, broader controls could backfire harder.

The real competition is diffusion, not frontier models. FPRI analysis: the US is pursuing "acceleration" — AI-first warfighting, military pace-setting projects. China is pursuing "diffusion" — narrower investments across manufacturing and consumer sectors, betting that AI tailored to specific industries will deliver economic benefits faster than frontier models.

4. AI Can't Fix What's Actually Broken (Eurasia Group, CNN)

China's real crisis is deflation, property collapse, and cratering consumer demand. You can't AI your way out of that.

Eurasia Group ranked China's deflation trap as its #7 global risk for 2026. Their assessment: Beijing will prioritize political control and technological supremacy over the consumption stimulus needed to break the cycle. Home prices have been falling for four and a half years — household wealth destruction on par with America's 2008 crash. Consumer confidence, investment, and domestic demand have all cratered.

The property sector used to be a quarter of GDP. Now it's only a fifth year. New housing starts are down 75% from their peak. Beijing bet that high-tech manufacturing would fill the gap, but state-driven investment has created overcapacity, and weak domestic demand means there aren't enough buyers to absorb it. Sound familiar? It's the same playbook that produced ghost cities.

The five-year plan is a tech strategy for an economy that needs a consumption strategy. CNN reported that China set its lowest growth target in decades while facing a "grave and complex landscape." Despite claiming to want to rebalance toward consumption, concrete policy plans remain timid. You can integrate AI into 90% of the economy, but if nobody's buying what the economy produces, you've just automated a deflation spiral.

Where This Lands

China just published a 52-mention AI plan while the US is focused on a war in the Middle East. The plan is ambitious, possibly overambitious — the VC money isn't there, the property sector is in freefall, and 90% AI integration is more slogan than strategy. But the direction is unmistakable: semiconductor self-sufficiency, AI diffusion across the entire economy, and a tech stack that doesn't need American chips.

Sources