Nvidia Signs $6 Billion Deal With AI Startup Poolside

Nvidia struck a $6 billion deal with Poolside to build an open-weight AI model competing with Chinese systems and hire over 100 staff.

maisiekooc
Maisie Morrison

AgentLocker Editor

AI News
Nvidia Signs $6 Billion Deal With AI Startup Poolside

Nvidia has agreed to a deal worth $6 billion with the artificial intelligence startup Poolside. The agreement was announced this week and covers technology licensing along with hiring staff.

As part of the deal, Nvidia will pay $6 billion to license Poolside's technology. The company will also invest $1 billion directly into Poolside at a $12 billion valuation.

More than 100 Poolside employees will move to Nvidia. These workers will join Nvidia's Nemotron project, which builds open-weight AI models.

What Poolside Brings to Nvidia

Poolside was founded in 2023 by Eiso Kant and Jason Warner. Warner previously served as chief technology officer at GitHub.

The startup sent a letter to shareholders explaining the deal. In the letter, Poolside said it wants artificial general intelligence to be built in the open rather than controlled by a small number of companies.

Poolside's founders, along with executive Margarida Garcia, will not join Nvidia. They plan to continue working on separate research projects.

The deal follows a difficult stretch for Poolside. The company said it missed a six week window last year to raise $2 billion needed for a large computing cluster.

Poolside said it later realized it would run out of computing power without new funding. The startup said Nvidia offered fewer limits on cash and infrastructure.

Nvidia's Push Into Open Models

Nvidia is best known for designing computer chips that power AI systems. In recent months, the company has also worked to build its own AI models.

Nvidia CEO Jensen Huang has supported open-weight models for years. These models can be downloaded for free and changed by developers for their own use.

Most major U.S. AI companies, including OpenAI and Anthropic, have focused on closed models instead. Closed models keep their code and data private.

In March, Nvidia launched a group called the Nemotron Coalition. Members include Mistral, Thinking Machines Lab and Perplexity, and they share data and computing resources.

Last month, Huang wrote a public letter about open AI models. He said success in AI will depend on whether the United States builds an open system that spreads across many industries.

Chinese companies have released several open-weight AI models in recent years. These include DeepSeek, Kimi and models from Z.AI.

When DeepSeek released a low-cost open model last year, it caused a drop in U.S. tech stocks. Investors worried Chinese AI technology was catching up to American systems.

Nvidia is now working on a large version of its Nemotron model. Reports say it could include more than a trillion training parameters.

Earlier this month, Nvidia released a smaller version called Nemotron 3.5 Lightning. The company is still building its largest model.

The Poolside deal also places Nvidia in competition with some of its own customers. Companies like OpenAI and Anthropic buy Nvidia chips but build closed models.

Poolside's most recent model, called Laguna S, has become popular among Western open-weight AI models. Nvidia's offer came around the same time Poolside released this model.

From our research desk
AI Jobs Automation Index
Which jobs are AI tools targeting most? We mapped 3,400+ AI tools to real job functions — with BLS employment & salary data.
Explore the index
maisiekooc

Written by

Maisie is a news writer at Agent Locker, covering the latest developments in artificial intelligence, emerging technology and the companies shaping the future.

Discover AI Agents