Nvidia expands support for Chinese AI models as U.S. restrictions loom

The chipmaker on Wednesday highlighted a "local AI initiative with optimizations for top open models," including DeepSeek's V4 Flash and Alibaba's Qwen 3.8, as well as systems from Google and Nvidia itself.

In a quarterly SEC filing, Nvidia said any "regulatory control or other restriction that limits our ability to provide products and services" supporting applications built on models "such as DeepSeek, Qwen or Kimi" could have a material impact on its business.

The move came as Chinese AI models have made capability leaps in 2026, and their growing adoption by developers worldwide, including in the U.S., has heightened concern among U.S. lawmakers.

Open AI models, the most capable of which are Chinese, have become a flashpoint in the U.S.-China technology rivalry. Unlike closed systems from OpenAI and Anthropic, open models can be modified and self-hosted.

A report on Thursday said the Trump administration is considering new tariffs on U.S. semiconductors. A White House official told CNBC that "unless officially announced by the Administration, any reporting about tariffs should be regarded as baseless speculation."

The U.S. is also mulling restrictions that would limit Nvidia's ability to support "third-party applications and models built on open-source foundation models originating in China," the company said.

Nvidia shares rose Thursday after the company reported better-than-expected second-quarter results and issued revenue guidance above estimates. Its stock is up more than 10% so far in 2026.

Nvidia is not alone in opposing government intervention in open-weight models. In July, Microsoft, Meta, Palantir and more than 20 other companies joined Nvidia in urging policymakers to avoid "premature restrictions" on such models.

An unnamed Nvidia employee told CNBC that "providing support for models worldwide allows developers to build on the American tech stack." The employee added that "developers using popular American and Chinese models will choose the stack that the model is optimized for, making the need to optimize for the American stack critical."

Chinese chipmakers, including Huawei and Alibaba, have announced their own optimizations for open models such as DeepSeek and Qwen in recent months.

"Developers will play a crucial role in building the winning AI ecosystem," the Nvidia employee said. "China has one of the largest populations of developers in the world, creating open-source foundation models. Every model should run best on the U.S. technology stack, encouraging nations worldwide to choose America."

Daniel Remler, a senior fellow at the Center for a New American Security, told CNBC that if Chinese AI technology "becomes the default for developing countries, those countries may be more likely to align themselves politically with Beijing and Chinese AI companies get a beachhead in their markets."

Charlie Dai, vice president and principal analyst at Forrester, said Nvidia's optimization for Chinese models reflects their "growing influence" within the global AI ecosystem. By optimizing for DeepSeek and Qwen, Nvidia is reinforcing its position as the preferred infrastructure layer, he said.

Nvidia announced in August it was optimizing hardware, including "day-zero support" on RTX GPU systems for Qwen3.8-27B. The company also said software updates made it "easy to cluster multiple DGX Spark systems together," which is important for running Chinese models like Z.ai's GLM 5.2 and DeepSeek V4 Flash.