Nvidia Licenses Poolside's AI Model Factory for $6 Billion

Nvidia has agreed to pay Poolside $6 billion to license its Model Factory software and hire 109 of its staff. The licence is non-exclusive, allowing Poolside to continue licensing the software elsewhere. This deal marks the third time Nvidia has used this structure, having previously paid Groq $20 billion for its inference technology and Enfabrica about $900 million.

The Model Factory is a system that Poolside used to build AI models, including Laguna, which was trained on Nvidia server chips. The founders of Poolside will remain with the company after the deal. In contrast to other companies like OpenAI, where Nvidia has invested heavily and reduced its commitment by half this month, Poolside intends to distribute the $6 billion to its investors by the end of next year.

Poolside began as a developer of a coding AI agent and moved into data centres before releasing its own open-source model on Nvidia server chips. The company's letter to investors explains why it stopped building frontier models, citing the rapidly increasing capital requirements and physical data centre space constraints. The founders believe that human-level capability will be fully commoditized by open source models, while superintelligence will not.