It’s the new battleground for China v the US. But Beijing has one big advantage

Although we are in the very early days of the artificial intelligence world, the US and China are emerging as the dominant forces. This will be a continuation of their geopolitical contest by other means.

We can assume that the two countries attain similar levels of tech prowess. Why? Because of the general and the specific. In general, China has caught up to, or surpassed, the US in every field of technology that it has designated as a priority.

China now leads the US in 69 of the 74 technologies classed as critical and advanced in the tracker maintained by the Australian Strategic Policy Institute, last updated in June.

It’s “a clear signal of a structural shift under way in global technological power”, according to ASPI. And when China’s firms industrialise technologies, they have a record of undercutting and overtaking US competitors.

The New York Times headline from July 15 tells the tale of one sector: “The American EV has been crushed”. China’s electric vehicles accounted for 60 per cent of all sales worldwide last year, according to the International Energy Agency.

And a CNN headline from last week tells another: “China’s humanoid robots have been taking over the global market. Now the US is banning them.”

China boasts 140 manufacturers turning out more than 330 models of humanoid robots, accounting for 90 per cent of all production globally last year “while US competitors like Tesla and Figure AI have struggled to start mass manufacturing”, as CNN put it.

And then there’s the specific. In AI specifically, the US has suffered serial rude shocks as one Chinese firm after another has released AI models that rival or exceed their frontier US counterparts.

The first such moment came last year when China’s DeepSeek unveiled its V3 large language model. Its performance is similar to OpenAI’s GPT-4o but developed at a reported cost of $US5.58 million ($7.96 million) compared with more than $US100 million for GPT-4. And with a fraction as many computer chips.

The second shock hit Wall Street hard a couple of weeks ago. Chinese start-up Moonshot published its Kimi K3 model. It’s rated as the world’s third most intelligent AI product, just behind Anthropic’s Claude Fable 5, according to the benchmarking site Artificial Analysis. But you can operate it for one-third of the cost.

The Chinese products are consistently cheaper. This hits the US industry hard. They cannot compete on price.

As a consumer simply using the AI function on your computer browser, you’re probably not conscious of the costs. But for companies and institutions that use AI on a large scale, costs have become prohibitive.

“Fed up with ballooning costs, companies big and small are starting to use lower-priced models, including some built in China,” as The Wall Street Journal reported last month.

Fortune magazine carried this startling headline in May: “Microsoft reports are exposing AI’s real cost problem: Using the tech is more expensive than paying human employees.”

And most businesses don’t need the top-end frontier products. As the paper reported one US tech entrepreneur, Mike Saeks, as saying: “It’s like driving a Lamborghini to go to the grocery store to pick up milk when that was designed to be raced around a track.”

The other key fact to know about the Chinese AI offerings is that they are all open to modification by users, so-called “open weight”, so that anyone can download them and customise them. The US products, by contrast, are closed and unalterable.

So much for the tech itself. While it’s essential to the competition, it’s only one element in a much bigger structure. For instance, and to continue the motoring metaphor, your car is essential for driving but useless without a system of roads, service stations and road rules.

The shock arrivals of top-notch Chinese AI models are treated as one-off events. “They are anything but,” says an American business adviser, Dewardric McNeal of Longview Global, writing for the US financial news site CNBC.

“Viewed collectively, they reveal something far more consequential than the emergence of several successful Chinese AI companies. They demonstrate that China has cultivated a frontier AI ecosystem capable of repeatedly producing world-class capabilities across multiple firms.” So it’s a systemic challenge to the US, not an episodic one.

What about the macro cost to each country? The US is betting its economic house on AI. There’s the investment in physical assets. In the past three years, just four of the big tech firms – Google, Amazon, Microsoft and Meta – have invested a combined $US1.1 trillion. They plan to spend another $US745 billion this financial year.

And the total planned investment for the entire American AI-related sector is estimated at perhaps $US9 trillion over the next four years. For perspective, total US economic output last year was $US30 trillion.

But above and beyond is Wall Street’s speculative frenzy of betting on the companies making these vast investments. The total market capitalisation of AI-related companies in the US today is $US27 trillion (the Chinese equivalent is just $US4 trillion). That’s one industry among dozens but priced as 40 per cent of the entire sharemarket.

The Bank of International Settlements, which is the co-ordinating body for the world’s central banks, offered some “instructive parallels” from history in its annual report in June.

“The canal mania of the 1830s, the British railway mania in the 1840s, the electrification exuberance of the late 1920s (roaring ’20s) and the dotcom boom of the late ’90s all shared one common trait: a genuine technological breakthrough that attracted capital in excess of what commercial returns could ultimately justify. These episodes ended with an eventual reversal in investment, inducing economy-wide recessions.” The current frenzied Wall Street expectations of AI companies “bear resemblance to these precedents”.

The coming bust is inevitable. Especially when these companies are uncompetitive compared to their Chinese competitors. And how much are these Chinese firms spending on AI? According to Stanford University’s AI Index, just $US12.4 billion last year.

US investment bank Goldman Sachs rates the Chinese sector a “buy” – it has only 10 per cent of global AI capitalisation but 16 per cent of global revenue. But while Wall Street faces a mighty reckoning and possible recession, China is not getting away cost-free.

One big reason its market is subdued and its economy flat is that Xi Jinping deliberately repressed China’s tech and entrepreneurial sector and its real estate markets five years ago. They have not fully recovered.

Why would he do such a thing? He was acting to pre-empt the damage that a convulsive boom and bust could do to China’s economy. He saw it as a major vulnerability, especially in the event of a war with the US.

So China already has paid a big price. America’s lies ahead.

Peter Hartcher is both international and political editor. His political column appears on Saturdays.

Get a weekly wrap of views that will challenge, champion and inform your own. Sign up for our Opinion newsletter.