Nvidia is pouring engineering resources into making its hardware the best place to run Chinese open-source AI models — even as the company warns in a regulatory filing that Washington could soon restrict exactly that kind of support.
The chip giant this week highlighted a "local AI initiative with optimizations for top open models," including DeepSeek's V4 Flash and Alibaba's Qwen 3.8, alongside systems developed by Google and Nvidia itself, CNBC reported on Thursday. The initiative aims to ensure that the most popular open models — the most capable of which now overwhelmingly come from Chinese labs — run best on Nvidia's stack rather than on the domestic Chinese alternatives rapidly maturing behind Huawei's Ascend processors.
At the same time, an SEC filing tied to Nvidia's second-quarter earnings warned investors that potential US government restrictions on Chinese-developed AI could have a material impact on its business. The dual message — embrace Chinese models commercially, brace for political blowback — captures the box Nvidia finds itself in as the US-China technology race intensifies.
For readers following AI industry coverage, the initiative is the clearest sign yet that the battle over open-weight AI has moved from debate to infrastructure.
Why Nvidia Is Optimizing for Its Rivals' Models
The strategy is commercially logical even if it looks counterintuitive. Open models can be downloaded, modified and self-hosted, unlike the closed systems sold by OpenAI and Anthropic — and whichever hardware a model is optimized for tends to win that developer's next deployment.
"Providing support for models worldwide allows developers to build on the American tech stack," an Nvidia employee who was not authorized to speak on the record told CNBC. "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."
The same employee pointed to the scale of Chinese open-source development: "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."
Nvidia's recent product moves back up the messaging. In August the company announced "day-zero support" for Alibaba's Qwen3.8-27B on its RTX GPU systems, and said updates making it easy to cluster multiple DGX Spark systems together were important for running Chinese models like Z.ai's GLM 5.2 and DeepSeek's V4 Flash.
Washington's Warning Shot
The commercial push runs directly into an unresolved policy question. Nvidia's SEC filing warned that the US government is mulling restrictions that would limit the company's ability to support "third-party applications and models built on open-source foundation models originating in China." Any "regulatory control or other restriction that limits our ability to provide products and services" supporting apps built on models "such as DeepSeek, Qwen or Kimi," the company warned, could materially affect its business.
US lawmakers have grown increasingly anxious about the adoption of Chinese models by American companies, which have made leaps in capability in 2026 while typically pricing below American alternatives. Access to open models has become a flashpoint in the wider race for technological supremacy between Washington and Beijing.
Tensions escalated further on Thursday, when a report emerged that the Trump administration is considering new tariffs on US semiconductors. A White House official told CNBC that "unless officially announced by the Administration, any reporting about tariffs should be regarded as baseless speculation."
Nvidia is not campaigning alone. In July, Microsoft, Meta, Palantir and more than 20 other companies joined the chipmaker in a statement urging policymakers to avoid "premature restrictions" on open-weight models — an unusually unified industry front that includes many of the same companies building closed models.
The Geopolitics of Default Infrastructure
Policy analysts warn that the stakes extend beyond any single company's revenue. "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," Daniel Remler, a senior fellow in the technology and national security program at the Center for a New American Security, told CNBC.
Forrester vice president Charlie Dai said Nvidia's optimization of platforms such as DeepSeek and Qwen reflects the "growing influence" of Chinese models within the global AI ecosystem — while simultaneously "reinforcing its position as the preferred infrastructure layer for AI developers and enterprises regardless of where leading models originate."
Chinese chipmakers are racing to do the same on their side. Huawei, which makes the Ascend processor series, and Alibaba have announced their own optimizations for open models including the DeepSeek and Qwen families in recent months, giving Chinese developers a fully domestic alternative stack.
A Test of the Open-Weight Coalition
Markets, for now, are rewarding the balancing act. Nvidia shares rose Thursday after better-than-expected second-quarter results and revenue guidance that topped estimates, and the stock is up more than 10% in 2026, according to CNBC.
But the company's SEC warning makes clear that its Chinese-model strategy survives only at Washington's pleasure. The coming months will test whether the open-weight coalition's lobbying can hold off restrictions — or whether the most consequential AI infrastructure decision of 2026 gets made in a regulatory filing rather than a product launch.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
