Trump's Nvidia Chip Deal with China Fizzles Amid Deep Distrust

When President 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 amid increasingly fragile geopolitical relations. It would provide a significant boost to China’s artificial intelligence ambitions while delivering a victory for the leading U.S. chip maker.

Cutting-edge artificial intelligence systems rely on staggering amounts of computing power, and Nvidia’s chips are widely considered the global gold standard. Competitors in China have yet to produce anything that rivals Nvidia’s best, and Trump’s decision undercut years of U.S. policy aimed at keeping such chips out of Chinese hands.

Former Biden and Trump administration officials have warned that the move could squander the U.S. AI industry’s lead over Chinese rivals by helping China close the gap until its own chip manufacturers can catch up to Nvidia.

But six months later, Beijing has not allowed any of its companies to buy even a single chip.

That stalemate exposes the depth of distrust between the world’s two tech superpowers. For decades, U.S. and Chinese companies worked together to create industry-disrupting products such as the iPhone. But over the past decade, that relationship has soured as both governments have begun to view technology as a linchpin of economic dominance.

Nvidia finds itself caught in the middle. The chip maker has become the world’s most valuable company by making its semiconductors indispensable to running AI systems. But Washington and Beijing increasingly view the technology as a national security issue, especially after witnessing how it has been used to coordinate attacks in Gaza, Ukraine, Venezuela and Iran.

Instead of turning to Nvidia, Chinese officials have pushed domestic companies to adopt local alternatives from chip makers such as Huawei and Cambricon. Trump said after meeting with Chinese leader Xi Jinping in Beijing last week that China lacked interest in the H200 chip he had approved, partly because of that push for industrial autonomy.

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

China’s AI companies are desperate for more computing power. Kevin Xu, founder of the hedge fund Interlinked Capital, which invests in AI technology, spent nine days in China recently meeting with leading AI startups. He said that every one of them pointed to a lack of computing power as the biggest obstacle to their development.

But Chinese companies are now starting to build their AI systems around these limitations rather than waiting for them to ease. Beijing wants them to continue to do so.

“It’s a balancing act,” Xu said. The government knows that companies need more computing power, but it also wants to push domestic chip makers to get better and faster.

Jensen Huang, Nvidia’s CEO, who traveled to Beijing last week with Trump’s delegation, said he did not mention the H200 chip to Chinese officials. But given China’s huge appetite for AI chips, he remains optimistic that the market will eventually open.

“The Chinese government has to make a decision — how much of the domestic market is protected, and how much is expanded by more AI capability,” Huang said on Bloomberg Television’s “The David Rubenstein Show” on Monday. “My sense is that the market will open over time.”

In March, Xi proposed an ambitious plan to deepen a technological self-sufficiency drive that China has pursued for a decade. Over the next five years, he wants the country to pursue breakthroughs in fields such as artificial intelligence, quantum computing and fusion energy.

That quest for self-reliance is becoming more feasible as domestic chip makers rapidly improve their products. Chinese tech giants such as Huawei and startups like Cambricon now produce chips that perform on par with the H200.

DeepSeek, a startup that has become a symbol of China’s growing AI prowess, has also become a symbol of the country’s move away from Nvidia.

When it released its latest AI model last month, DeepSeek 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 the country’s long-standing effort to develop advanced technology domestically.

While Washington and Beijing have effectively blocked Nvidia’s most advanced products from entering the world’s largest chip market, the restrictions have done little to slow the chip maker’s business. On Wednesday, Nvidia reported quarterly profits of $58.3 billion. The company said it had no sales of its signature AI chips in China and did not expect that to change in the short term.

While domestic chips are used for some tasks now, Chinese tech companies still rely on Nvidia chips for model training — the grueling process of teaching AI models how to run. Part of the reason is that Chinese chip makers have struggled to produce advanced chips in sufficient quantities to meet demand.

Many Chinese AI companies are using a workaround to avoid buying their own Nvidia chips. They are renting remote access to chips housed in data centers operated by other companies, often offshore. When two AI startups, Minimax and Zhipu AI, went public earlier this year, their disclosures showed that the companies were spending multiples of their revenue to train their models using “cloud services.”