More than 100 Poolside engineers to join Nvidia’s Nemotron project
Nvidia will pay $6 billion to license artificial intelligence startup Poolside’s technology and hire more than 100 of its engineers to build an open-weight AI model aimed at competing with Chinese rivals including DeepSeek and Kimi K3, according to people familiar with the matter. The chip maker also will invest $1 billion in Poolside at a $12 billion pre-money valuation. The acquired engineers will join Nvidia’s Nemotron project, where they will work on its largest and most sophisticated open-weight models under development, the people said.
The deal adds to Nvidia’s effort to position open-weight AI models — whose source code and numerical weights are publicly available — as a U.S. alternative to Chinese open-weight systems that have trailed frontier U.S. AI capabilities by only a few months.
Poolside announced the restructuring in a letter to shareholders on Thursday that was reviewed by The Wall Street Journal. The company was founded in 2023 by Eiso Kant, a software developer, and Jason Warner, the former chief technology officer of GitHub. Poolside’s leadership — including Kant, Warner and operations executive Margarida Garcia — will not join Nvidia and will continue working on unspecified research projects, according to the letter.
In the letter, Poolside’s founders wrote that the Nvidia deal was meant to secure a future in which artificial general intelligence, or humanlike computer reasoning, “would not be a closed technology controlled by few but one built by many out in the open.” They said Nvidia was the right partner in part because the chip maker has fewer computing-infrastructure or cash limitations than Poolside did.
The restructuring followed a failed fundraising effort late last year. “At the end of last year, we had a 6 week window in which to raise $2 billion dollars to pay for a 40,000 GB300 cluster coming online in January,” Poolside’s letter read, referring to a common type of Nvidia server used in model training. “We didn’t close it in time, and we lost the cluster.” Poolside said the company would run out of computing power as soon as next year without access to the capital and data centers it needed.
Poolside’s most recent model, known as Laguna S, had emerged as one of the most popular open-weight models in the West before the deal was struck. Although Poolside and Nvidia have a longstanding relationship, the agreement came together in the past few weeks. Nvidia made its offer to Poolside around the same time the startup released Laguna S.
Nvidia’s Nemotron project debuted in 2023 as the chip maker’s effort to build open-weight models. Earlier this month, Nvidia released Nemotron 3.5 Lightning, a “lightweight” version of its open-weight model, and is developing what is rumored to be its largest version yet, with more than a trillion training parameters — putting it in the same league as some of the largest AI models ever designed, though some Chinese open-weight models are larger still.
The U.S.-China AI competition has sharpened since DeepSeek released its first low-cost, open-weight model in January 2025. The release caused a sell-off in U.S. tech stocks because investors interpreted it as proof that Chinese models were on track to match or surpass American AI technology. Venture capitalist Marc Andreessen called it a “Sputnik moment” for AI. New Chinese players have since emerged with open models that trail frontier U.S. AI capabilities by only a few months, including Moonshot’s Kimi and Z.AI’s GLM family.
Nvidia has moved to build out its open-weight ecosystem in response. In March, the company launched the Nemotron Coalition, an association of leading open-model developers including Mistral, Thinking Machines Lab and Perplexity. The same month, the Journal reported that Reflection AI, another Nvidia-backed startup that investors have described as “DeepSeek of the West” for its sophisticated open-weight models, was in advanced talks to raise $2.5 billion at a $25 billion valuation.
Last month, Nvidia Chief Executive Jensen Huang published his first-ever post on the social-media network X, a letter titled “Open Weights and American AI Leadership.” Huang argued that whether the U.S. wins the AI race “will be judged not by one frontier AI model, but by whether the United States builds a strong, open ecosystem that diffuses into every sector.”
The Poolside deal also places Nvidia in competition with some of its biggest chip customers. OpenAI and Anthropic are among the largest purchasers of Nvidia processors, and several major AI labs — including OpenAI, Anthropic and the research arm of Alphabet’s Google — have begun designing their own custom silicon to power their models. Most of those frontier labs have committed far more resources to proprietary, “closed” AI models than to open-weight alternatives, the Journal reported.
Newcomer, an independent tech newsletter, first reported the Poolside licensing and investment deal.