Nvidia agrees to pay Poolside $6 billion for AI model licenses, hiring more than 100 people

Nvidia has agreed to pay Poolside $6 billion to license Poolside’s artificial intelligence models, according to people familiar with the situation. The agreement is part of a broader push by large AI platform companies to secure model capabilities without having to replicate every breakthrough in-house, an approach that has become increasingly common across the industry.

The deal also includes plans to extend job offers to more than 100 employees from Poolside. Hiring startup talent is often a key pathway for acquirers or strategic partners to preserve technical know-how and accelerate integration, particularly when the value at stake is tied to specific model performance and training methods.

Poolside’s contribution in the arrangement is described in the reporting as its AI models. In general terms, AI models are systems trained on large data sets to perform tasks such as generating text or producing other outputs, and licensing them gives a buyer rights to use the models for particular purposes rather than building an equivalent model from scratch.

For Nvidia, the licensing structure reflects its position at the center of the AI stack through its hardware and software ecosystem. Nvidia’s business depends on developers running and training AI workloads on its platforms, so expanding the set of usable models and tooling can help keep customers building on its infrastructure.

Still, the reporting does not specify how the licenses would be used, what exact models are covered, or whether the licensing terms include exclusivity, transferability, or certain performance benchmarks. It also does not disclose whether the $6 billion figure reflects a single payment, a multi-year schedule, or a mix of upfront and milestone-based components.

The news comes as companies in the AI sector continue to compete on access to high-performing models and the talent behind them. Poolside’s reported staff move suggests Nvidia expects meaningful technical overlap between the startup’s model work and Nvidia’s longer-term strategy for deploying AI at scale.

For executives and engineers, the immediate operational question will be how quickly the models can be integrated into Nvidia’s existing developer and enterprise offerings. The timeline for that integration, along with any technical handoff details, remains unclear from the available reporting.

What to watch next is whether Nvidia and Poolside provide additional confirmation and details, such as the specific model families included, the geography and use-case boundaries of the licensing agreement, and the role the newly hired staff will take once they join. Until then, many of the practical implications of the $6 billion figure hinge on terms that have not been publicly described.