$NVDA is making a serious pivot into open-weight AI models — and it's not subtle.
They're paying $6 billion to license tech from Poolside, bringing over 100 of their engineers onto the Nemotron team. On top of that, they're investing $1 billion at a $12 billion valuation. Poolside stays independent, but the collaboration is deep.
The goal? Build a competitive open-weight U.S. model that can go head-to-head with Chinese systems like DeepSeek and Kimi — while staying more flexible and accessible than closed models from OpenAI or Anthropic.
This isn't just about selling chips anymore. Nvidia is moving upstream, shaping the AI software stack itself. It's a structural shift in how they position themselves in the AI supply chain — and it signals where they think the leverage is heading next.
They're paying $6 billion to license tech from Poolside, bringing over 100 of their engineers onto the Nemotron team. On top of that, they're investing $1 billion at a $12 billion valuation. Poolside stays independent, but the collaboration is deep.
The goal? Build a competitive open-weight U.S. model that can go head-to-head with Chinese systems like DeepSeek and Kimi — while staying more flexible and accessible than closed models from OpenAI or Anthropic.
This isn't just about selling chips anymore. Nvidia is moving upstream, shaping the AI software stack itself. It's a structural shift in how they position themselves in the AI supply chain — and it signals where they think the leverage is heading next.
