Nvidia is preparing to restart sales of its H20 artificial intelligence chips to China after a turbulent stretch of policy shifts in Washington. The company said Monday that it is filing applications and expects to receive U.S. government licenses soon, with deliveries beginning shortly after.
The decision matters because the H20 chip has become a focal point in the U.S.-China technology standoff. It is not Nvidia’s most advanced AI processor, but under existing export controls it is the most powerful chip Nvidia can legally sell to China.
What Nvidia Says Is Coming Next
According to Nvidia, the company is moving ahead with license applications that would allow H20 chip sales to resume in China. The company expects those licenses to arrive soon, then plans to begin deliveries shortly after approval.
Nvidia is also introducing a new chip for the Chinese market called “RTX Pro.” The company described that chip as “fully compliant” with regulations and positioned it for digital manufacturing uses, including smart factories and logistics.
Together, the H20 restart and the new RTX Pro chip show how Nvidia is trying to keep serving a major market while staying inside changing regulatory limits. The company is not simply selling the same products everywhere. It is shaping products around what U.S. rules allow.
Why The H20 Chip Matters
The H20 is designed for “inference” work. In plain terms, that means it is used to run existing AI models in everyday applications rather than to train new AI systems from the ground up.
That distinction is important because the chip sits between commercial demand and export control policy. The H20 is less powerful than Nvidia’s top AI processors, but it remains attractive to Chinese companies because of its memory bandwidth and Nvidia’s software ecosystem.
Chinese tech giants including ByteDance, Alibaba, and Tencent had been stockpiling these chips in the first three months of this year. The source article says that buying happened in anticipation of stricter export controls.
The appeal was not only raw hardware. Nvidia’s widely used software ecosystem can make its chips easier to deploy, which gives the company a practical advantage even when Chinese alternatives are available.
How The Restrictions Shifted
The regulatory back-and-forth began in April, when the Trump administration restricted H20 sales. The source article says that move could have cost Nvidia $15 billion to $16 billion in revenue, based on how much Chinese firms reportedly spent on the chips in the first quarter alone.
The restriction targeted chips above certain performance thresholds. Those thresholds included total memory bandwidth of 1,400 gigabytes per second or input/output bandwidth of 1,100 GB per second.
But the ban did not stay in place for long. Soon after CEO Jensen Huang attended a $1 million-per-head dinner at Trump’s Mar-a-Lago resort in early April, the administration paused the restriction.
According to NPR, the White House changed course after Nvidia promised new U.S. data center investments. Within a week of NPR’s report being published, Nvidia announced plans to build AI servers in the U.S. worth as much as $500 billion over the next four years, with help from partners such as TSMC.
Why Lawmakers Are Pushing Back
The reversal has drawn criticism from U.S. lawmakers. Their concern is that policy changes like this could weaken U.S. efforts to limit China’s AI capabilities.
The debate is partly shaped by DeepSeek. The Chinese startup drew major attention earlier this year after building an impressive model using Nvidia’s H800 chips, which the source article describes as slightly more powerful predecessors to the H20.
The U.S. banned sales of those H800 chips back in October 2023. Even so, Chinese suppliers have managed to find workarounds, according to the source article.
That history is why the H20 has become more than a product story. For critics, it is evidence that export controls can be difficult to enforce and politically vulnerable. For Nvidia, it is a business line caught in the middle of rules that can change quickly.
A Commercial And Political Balancing Act
Nvidia spokesman Hector Marinez told TechCrunch that Huang has been meeting with officials in Washington and Beijing this month and “emphasizing the benefits that AI will bring to business and society worldwide.”
That message reflects Nvidia’s position: it wants to keep selling AI hardware while presenting AI as broadly useful for business and society. At the same time, U.S. policymakers are weighing national security concerns against the commercial interests of a major technology company.
The current episode shows how unstable that balance can be. Restrictions were imposed, then paused. Nvidia warned of major revenue exposure, then moved toward renewed deliveries. Lawmakers objected, while the company prepared both license applications and a new China-specific chip.
For the AI industry, the lesson is straightforward: export controls are now part of product strategy. Chips, markets, government licenses, and diplomatic pressure are no longer separate tracks. In Nvidia’s China business, they are moving together.