Nvidia expands support for Chinese AI models as U.S. restriction risk threatens business
Growing adoption of Chinese open AI models is sharpening competitive and regulatory pressure for U.S. chipmakers. Nvidia is optimizing its hardware for systems such as DeepSeek and Qwen while warning that possible U.S. limits on China-originating models could materially affect its business.
Highlights
- Nvidia expanded support for Chinese open AI models like DeepSeek V4 Flash and Alibaba Qwen 3.8 to address regulatory risks and strengthen its market position.
- Nvidia warned in an SEC filing that new U.S. restrictions limiting support for Chinese-developed open-source foundation models could materially impact its business.
- Nvidia's stock rose after strong Q2 results and upward revenue guidance, with shares up over 10% in 2026 amid industry opposition to tighter U.S. controls on open models.
Model support strategy and regulatory warning
As reported by CNBC, Nvidia on Wednesday highlighted a local AI initiative with optimizations for leading open models, including DeepSeek’s V4 Flash and Alibaba’s Qwen 3.8, alongside systems developed by Google and Nvidia itself.The move reflects Nvidia’s effort to strengthen support for Chinese artificial intelligence models before domestic chipmakers gain an edge in tailoring hardware for those systems. At the same time, the company said in an SEC filing tied to its second-quarter earnings that potential White House restrictions on AI developed in China pose a business risk.
Nvidia said the U.S. is also considering restrictions that would limit its ability to support third-party applications and models built on open-source foundation models originating in China. The company warned that any regulatory control or other restriction limiting its ability to provide products and services for applications built on models such as DeepSeek, Qwen or Kimi could have a material impact on business.
Market reaction and wider industry stakes
Chinese AI models are making rapid gains in capability in 2026, and their growing use by developers globally, including in the U.S., is intensifying concern among lawmakers in Washington. Access to open AI models has become a key flashpoint in the technology rivalry between the U.S. and China because such systems can be modified and self-hosted, unlike closed models from OpenAI and Anthropic.Political pressure is also building on the semiconductor side. A report emerged on Thursday that President Donald Trump’s administration is considering new tariffs on U.S. semiconductors, although a White House official told CNBC that any reporting on tariffs should be treated as baseless speculation unless officially announced by the administration.
Nvidia shares rose on Thursday after the company posted better-than-expected second-quarter results and gave revenue guidance above estimates. The stock is up more than 10% so far in 2026, and the company is not alone in opposing tighter U.S. intervention in open models, with Microsoft, Meta, Palantir and more than 20 other companies joining a July statement urging policymakers to avoid premature restrictions.
Our earlier coverage of proposed U.S. semiconductor tariffs explained that the Trump administration was weighing a new round of duties that could extend beyond chips to a wider range of tech products such as laptops, servers and gaming hardware. We noted that the plan was still in an early stage and could be rolled out in phases, while also linking any relief to commitments to invest in U.S. semiconductor manufacturing.