TAIPEI, Taiwan — When President Donald Trump announced late last year that Nvidia could sell one of its most powerful chips to China, the deal looked like a rare win-win in a fraying geopolitical relationship. It would provide a major boost for China’s artificial intelligence ambitions, while handing a win to America’s leading chipmaker.

Cutting-edge AI systems run on staggering amounts of computing power, and Nvidia’s chips are considered the gold standard worldwide. Chinese competitors have yet to build anything that rivals Nvidia’s best, and Trump’s decision undercut years of U.S. policy designed to keep those chips out of China’s reach.

Former Biden and Trump administration officials warned the move could squander the lead that U.S. AI companies held over Chinese rivals, by helping China close the gap until its own chipmakers could catch up to Nvidia.

But six months on, Beijing has not allowed any of its companies to buy a single one.

The impasse lays bare the depth of the mistrust between the world’s technological superpowers. For decades, U.S. and Chinese companies worked side by side to create products like the iPhone that upended industries. But the relationship has soured over the past decade as both governments came to see technology as the fulcrum of economic supremacy.

Nvidia is trapped in the middle. The chipmaker became the world’s most valuable company by making its semiconductors indispensable to running AI systems. But Washington and Beijing increasingly view that technology as a matter of national security, especially after watching how it has been used to coordinate attacks in the Gaza Strip, Ukraine, Venezuela and Iran.

Rather than turn to Nvidia, Chinese officials have pushed domestic companies toward homegrown alternatives from chipmakers such as Huawei and Cambricon. After meeting last week in Beijing with President Xi Jinping, Trump said China’s lack of interest in the chip he had approved, known as the H200, had been driven in part by this push for industrial self-reliance.

“They chose not to. They want to try to develop their own,” he said.

Chinese AI companies are hungry for more computing power. Kevin Xu, founder of Interconnected Capital, a hedge fund that invests in AI technologies, recently spent nine days in China meeting with leading AI startups. Every one of them, he said, pointed to the lack of computing power as the biggest factor holding them back.

But Chinese firms are starting to build their AI systems around those constraints rather than waiting for them to ease. And Beijing wants them to keep doing that.

Jensen Huang, Nvidia’s CEO, who traveled to Beijing last week as part of Trump’s delegation, said he didn’t bring up the H200 chip with Chinese officials. But he remains optimistic that the market will eventually open, given China’s enormous appetite for AI chips.

“The Chinese government has to decide — how much of their local market do they want to protect and how much of their local market do they want to expand with more AI capacity,” Huang said.

In March, Xi laid out an ambitious plan to deepen China’s already decade-long push to rely on its own technology. Over the next five years, he wants the nation to pursue breakthroughs in fields like AI, quantum computing and fusion energy.

The drive for self-sufficiency is becoming more attainable as domestic chipmakers rapidly improve their offerings. Huawei, China’s tech giant, and Cambricon, a startup, and others are now producing chips that perform on par with the H200.

DeepSeek, the startup that has come to symbolize China’s rising AI capabilities, has also become a sign of the country’s shift away from Nvidia.

When it released its latest AI model last month, the company said for the first time that its new system had been optimized to run on chips made by Chinese tech giant Huawei — a small but meaningful milestone in China’s long-running effort to develop advanced technologies at home.

Although Washington and Beijing have effectively blocked Nvidia’s most advanced products from the world’s largest chip market, the restrictions have done little to slow the chipmaker’s business. On Wednesday, Nvidia reported a quarterly profit of $58.3 billion.

While Chinese tech companies now use domestic chips for some tasks, they still rely on Nvidia chips for training, the demanding process of teaching an AI model how to function. Part of the reason is that Chinese chipmakers have struggled to manufacture advanced chips at volumes sufficient to meet demand.

Many Chinese AI companies are renting remote access to chips housed in data centers operated by other companies, often outside the country. When two AI startups, MiniMax and Zhipu AI, went public in Hong Kong earlier this year, their disclosures showed the companies were spending several times their revenue to train their models using “cloud services.”

But this approach has drawbacks, said Jiang Tianjiao, associate professor at Fudan University in Shanghai. It means slower processing times and a higher risk of data leaks. It makes Chinese companies reliant on third-party data center operators, which could cut them off at any time. They also face the risk that the U.S. could pass regulations to close off remote access to Nvidia chips altogether.

Chinese companies are expected to spend $123 billion on AI chips and data centers this year, according to Bernstein Research. U.S. tech companies, by contrast, are expected to spend about $1 trillion.