NVIDIA has announced plans to resume sales of its H20 GPU in China, marking a reversal of a restriction imposed by the United States government earlier this year.
Founder and CEO Jensen Huang confirmed the development during customer briefings this month, stating that the company had begun filing export applications with the necessary authorities.
The original ban, enforced in April, cited national security concerns regarding the potential military use of advanced AI accelerators by Chinese entities. However, NVIDIA has now received assurance from US regulators that export licences will be granted for renewed H20 shipments.
As part of its localisation strategy, NVIDIA is also expected to launch a new AI chip for the Chinese market as early as September. The chip, based on the Blackwell RTX Pro 6000 architecture, will omit export-sensitive features such as high-bandwidth memory and NVLink interconnects to comply with current US regulations.
The chip has been designed specifically for industrial AI use cases, including digital twins for smart factory and logistics applications.
Speaking during a recent visit to Washington, D.C., Huang shared a broader vision of AI’s role in global infrastructure. He said: “The world has reached an inflection point – AI has become a fundamental resource, like energy, water, and the Internet.”
Huang reaffirmed NVIDIA’s support for open-source AI development and foundation models, highlighting the company’s ambition to enable innovation across emerging economies in regions such as Latin America, Europe, and Asia.
“General-purpose, open-source research and foundation models are the backbone of AI innovation,” he said. “We believe that every civil model should run best on the US technology stack, encouraging nations worldwide to choose America.”
