There Is No AI War Between China and the US

And believing there is may be the biggest distraction of our time — blinding us to the true forces shaping the future. 

Why the illusion of an AI race persists

In my keynote, I began with a simple question: “Who here uses ChatGPT weekly?” Almost all hands go up. Then: “Who here uses a Chinese AI model weekly?” Just a handful.

So the conclusion seems obvious: the US must be winning.

But when you look at the reality on the other side of the world, the picture changes instantly. Over 500 million Chinese use a local LLM every week. And NVIDIA’s Jensen Huang openly states that half of the world’s AI developers are in China — while many developers in the US are also Chinese.

So why do we still perceive that China is “behind”?

Because we see what comes out of China as the “artificial” part of AI, while we trust that the “Intelligence” part of AI is being built in the U.S. And we still use outdated assumptions that China cannot innovate, even though the evidence has completely changed. That cognitive gap is exactly what I explore in detail in China’s Next Miracle.

Why there is NO AI war: A versus I

We imagine a race because we assume China and the US are trying to run the same race. They aren’t. Their AI worlds could not be more different.

In my keynote, I explained it through a simple but powerful visual lens: A versus I. China’s AI strategy revolves around four A’s. The US strategy revolves around four I’s.

China’s AI model:

Accessible: China is doubling down on open-source. DeepSeek triggered a nationwide shift where almost every Chinese tech company now builds its models with open source and open collaboration. Open-source is not a side project in China — it is the national direction. China uses AI to maximise access, not restriction.

Affordable: Chinese companies know how to scale and optimise. Instead of overinvesting in giant models or hyperscale compute, China made AI cheap, efficient, and fast to deploy. Affordability is not an accident — it is a strategy to win the Global South, where most people cannot pay $20 per month for AI services from OpenAI or Microsoft. China sees AI as a lever to become the service factory of the non-Western world. 

Applicable: China’s focus is not on building one universal model like ChatGPT. China goes mostly vertical. Sector by sector. Business by business. Application by application. The country builds AI that solves real problems in healthcare, logistics, schools, factories, and government services. The AI hospital in Beijing, operating without doctors and diagnosing 20,000 patients per week, is one of the clearest examples of this approach. AI is not a product — it is an application engine.

Augmenting: China’s government views AI as a tool to strengthen society, not disrupt it. It invests heavily in AI governance — not to slow AI down, but to collectively build global trust so adoption accelerates. Beijing’s AI ambition is to augment humanity, not impose AI on it.

The US AI model:

IP-guarded: American AI companies build proprietary systems surrounded by deep moats. Their foundation models are protected, closed, and monetised through APIs. The strategy is clear: defend the intellectual property at all cost.

Investment-driven: American tech companies rely on locking us in with subscription models to recover billions in investment. The whole stockvalue of their company relies on future returns. AI became a business model, not a public good as it started off with OpenAI.

Infrastructure-focused: The US sees AI through the lens of chips, compute power, hyperscalers, and AGI. Performance is the priority. Bigger models, more GPUs, more data, more power. The goal is to build the most capable frontier models on earth.

Imposing: The official US AI Action Plan frames AI as a zero-sum game the US must win to preserve its global technological dominance. The approach is competitive, not collaborative, and explicitly prioritises American leadership over global governance.

Two models, two futures — no war

Once you understand the A’s and the I’s, the myth of a US-China “AI war” collapses instantly. The US and China are not competing for the same markets, the same users, or the same outcomes. One wants AGI and global dominance. The other wants mass adoption and global service capacity.

You cannot have a war when you are playing different games.

The real competition is not China versus the US. It is open-source versus proprietary. It is affordability versus monetisation. It is real applications versus AGI. It is collaboration versus competition. It is global governance versus worldwide monopoly.

And on each of these axes, China and the US are positioned on opposite sides.

Europe’s blind spot

Europe continues to look mainly toward the US, absorbing American narratives and ignoring Chinese realities. This is very dangerous. Because if Europe only chooses proprietary American AI, we risk becoming digitally colonised by systems we cannot control — something I warn about explicitly in China’s Next Miracle.

Meanwhile, China’s approach to regulation, governance, safety, and AI for public good is actually far closer to European values. But geopolitical distrust prevents forward-looking dialogue. Europe is staring at the wrong player once again.

China’s next miracle — and our choice

What I witnessed at the book launch was encouraging: people are beginning to appreciate that China is not copying the West anymore. It is flipping the innovation model — performing a global Fosbury-flop. AI is accelerating China’s transformation.

There is no war. There is only a profound divergence of models. And the future will be shaped by those who dare to see it clearly.

So the real question is the one I ended my keynote with:

Do we dare to jump differently? Do we dare to rethink innovation the way China is doing? Do we dare to take the red pill?

Because until AGI become a reality, the future will not be decided by who has the best or biggest AI model, but by who has the courage to flip the model completely.