Big tech companies are spending hundreds of billions developing artificial intelligence, and that spending has been floating the economy. But increasingly, the markets seem to think we're due for an unpleasant correction.

Just look at AI chipmaker Nvidia, which reported blockbuster earnings last week but whose shares then sank. It’s hard to say if the markets are headed for an AI bubble burst, but AI spending outside the U.S. may offer some hints.

The past few weeks have been choppy for Chinese tech companies, but for different reasons than U.S. ones.

There are multiple fronts in the AI cold war, including — of all things — milk. And sorry Wisconsin, China has the lead there.

“The state ordered, like, one of the state's biggest dairy companies to create a dairy- and milk-based, nutrition-focused LLM [large language model],” said Kendra Schaefer, head of tech policy research at the consulting group Trivium China.

LLMs are the technology that powers the likes of ChatGPT.

Robot-enhanced dairy operational efficiencies are probably not the most economically efficient use of resources, but it speaks to China’s broader strategy on AI: Fewer massive investments in data centers to enable robot superintelligence, more focus on having businesses and consumers actually use the robots right now.

“China's goal isn't to hit some kind of finish line or develop some kind of super technology, but its goal is to diffuse AI as widely as possible throughout the economy and society,” Schaefer said.

Chinese tech giant Alibaba recently announced a $50 billion AI investment over the next three years. By comparison, one project led by OpenAI is supposed to cost $500 billion over the next four years.

China’s approach is cheaper. “It may be less prone to a bubble because the money is not being dumped into that bet,” Schaefer said.

There’s another reason China isn’t going all in on data centers though. The U.S. isn’t letting it buy the fanciest AI chips to make building a new data center worth it.

“Obviously there are bubble concerns, but I think that's not what's holding back the Chinese development of equivalent infrastructure. It is the the constraints that they have on on some of the supplies and inputs into it,” said Chris McGuire, a senior fellow for China and emerging technologies at the Council on Foreign Relations.

If American Big Tech is right and demand for AI justifies all that spending, our economy will reap the benefits first, because it’s more closely tied to AI.

But if it doesn’t? Well, we may be in for a bubble burst.