1 week ago
Nvidia Expands AI Model Push With $6 Billion Poolside Deal
Nvidia is best known for making computer chips used for artificial intelligence.
Now it is also trying to build the AI models that run on those chips.
Nvidia has agreed to a $6 billion deal involving the startup Poolside.
More than 100 Poolside workers, mostly engineers, are expected to join Nvidia.
They will work on Nemotron, a group of AI models that users can download and change.
Nvidia is trying to compete with companies such as OpenAI, Anthropic and DeepSeek.
Poolside’s founders said the deal will give the startup access to the computing resources it needs.
The agreement is unusual because OpenAI and Anthropic are also major Nvidia customers.
At the same time, the computers needed to develop AI may become more expensive because memory chips are in high demand.
Nvidia will pay $6 billion to license Poolside’s technology and hire most of its engineers, according to The Wall Street Journal.
Nvidia plans to invest an additional $1 billion in Poolside at a pre-money valuation of $12 billion.
More than 100 Poolside employees are expected to join Nvidia’s Nemotron team, while senior Poolside leaders remain outside Nvidia.
The engineers will work on Nemotron, Nvidia’s family of open-weight AI models, as the company challenges OpenAI, Anthropic and DeepSeek.
The deal comes as AI server prices may rise more than 15%, partly because demand for memory chips is increasing.
- Who
- Nvidia, Poolside and more than 100 Poolside employees are involved; the deal affects competitors and customers including OpenAI and Anthropic.
- What
- Nvidia will license Poolside’s technology, hire most of its engineers and invest another $1 billion in the startup.
- Where
- The deal concerns Nvidia’s Nemotron organization and the broader US-China AI model competition.
- When
- The agreement was reported recently; the articles also refer to model competition that intensified after early 2025.
- Why
- Nvidia wants to expand from AI hardware into model development, strengthen its open-weight AI efforts and gain Poolside’s technology and engineering expertise.
Strategic expansion
Competitive tensions
Open-weight AI models
Strategic expansion
Open-weight models can be downloaded, adapted and modified, potentially giving businesses more flexibility and lowering operating costs.
Competitive tensions
Nvidia’s expansion into models could put it in competition with OpenAI and Anthropic, which are also major Nvidia customers.
Poolside’s future
Strategic expansion
Poolside’s founders said the agreement provides access to the computing resources needed for the startup’s longer-term ambitions.
Competitive tensions
Most Poolside engineers will leave for Nvidia, while the company’s senior leadership remains outside Nvidia, creating an unusual structure for the deal.
AI infrastructure costs
Strategic expansion
Nvidia’s broader model push could increase its influence over both the hardware and software used in AI development.
Competitive tensions
Rising memory-chip costs and expected server-price increases could make training and running AI models more expensive for customers.
Key facts
- Technology license
- Nvidia will pay $6 billion to license Poolside’s technology.
- Additional investment
- Nvidia plans to invest $1 billion in Poolside at a $12 billion pre-money valuation.
- Employees joining Nvidia
- More than 100 Poolside employees, mostly engineers, are expected to join Nvidia.
- Nvidia team
- The employees will work on Nemotron, Nvidia’s family of open-weight AI models.
- Poolside founding
- Poolside was founded in 2023 by Eiso Kant and Jason Warner.
- Potential server-price increase
- Some Nvidia customers were told prices for systems containing Nvidia AI chips could rise by more than 15% for early-next-year shipments.
- Memory-chip suppliers
- Samsung, SK Hynix and Micron dominate the DRAM market, while AI data-center demand is pressuring supply.





