The largest manufacturer of chips for artificial intelligence is betting on open models in the race for global technological leadership. At the center of attention is a deal with the AI startup Poolside, valued at $6 billion, which is intended to be a response to the rapid progress of Chinese large language models, including DeepSeek and Kimi K3.
A strategic move amid geopolitical pressure
The structure of the agreement with Poolside is telling: Nvidia is paying the startup $6 billion for licenses to AI models, integrating more than 100 employees into its team, and adding $1 billion as direct investment. This mirrors the recent acquisition of chip developer Groq, where $20 billion was paid for licenses and personnel. Such moves point to a systematic strategy: instead of acquisitions—licensing and talent, which preserves flexibility and accelerates development.
My analysis shows that Nvidia is deliberately moving away from the "closed ecosystem" model. The company is betting on open-weight solutions that developers can freely use and adapt. This is not just democratization of AI, but also a pragmatic calculation: open models stimulate demand for GPU infrastructure, where Nvidia remains a monopolist. Supporting open source is a long-term investment in its own market.
The challenge to OpenAI and the political context
The deal with Poolside is a direct blow to the positions of OpenAI and Anthropic. Models with open weights are significantly cheaper to operate and easier to customize, making them attractive to businesses. If Nvidia manages to create a competitive alternative, it could redistribute a significant share of the corporate AI solutions market.
Against this backdrop, discussions have intensified in the U.S. administration about possible restrictions on the use of Chinese open-weight models. Notably, 25 American companies, including Nvidia, Meta, and Microsoft, have opposed the bans, warning that such measures would only strengthen the position of closed developers and weaken U.S. technological leadership. The trigger for the debates was the release of Kimi K3 from Moonshot AI, which the White House accused of distilling Anthropic's technologies.
In my view, the situation is paradoxical: American giants are forced to defend openness in order not to lose the initiative to Chinese competitors, who have already proven the effectiveness of this model. Nvidia's deal with Poolside is not just a commercial move, but an attempt to seize the agenda and create a "national" open AI that can compete with the best examples from China. The only question is whether Nvidia will manage to implement this ambitious plan before regulators and the market change the rules of the game.