Nvidia has agreed to pay AI model-building startup Poolside $6 billion for a non-exclusive license to its model-development software, known as the Model Factory, while also investing $1 billion in the remaining company at a $12 billion pre-money valuation and extending job offers to 109 of its employees.

 

The terms were detailed in a letter Poolside sent to its investors and first reported by Newcomer. The three founders are remaining with the company, which has stated that the arrangement is neither an acquisition nor an acquihire. Poolside plans to distribute the $6 billion licensing proceeds to investors by the end of next year.

 

This structure mirrors previous Nvidia transactions with Groq and Enfabrica. In each case, Nvidia secured technology rights and talent through licensing and hiring while allowing the target company to continue operating independently. 

 

The Groq deal, valued at roughly $20 billion for technology and staff, left that company free to raise further capital, including a later round at a $3.5 billion valuation that included Nvidia participation.

 

Poolside began as a developer of coding-focused AI agents before expanding into data centers and releasing open-weight models trained on Nvidia hardware. Its Laguna family of models has been positioned as a Western open-weight alternative for software engineering tasks, including code generation, debugging and optimization. Nvidia was already an investor, having committed capital to Poolside in a prior round.

 

According to the investor letter, Poolside faced significant capital constraints late last year. The company described a six-week window in which it needed to raise $2 billion to secure a large cluster of Nvidia GB300 chips. It failed to close the financing in time and lost access to the hardware. The letter argued that next-generation frontier models will require clusters an order of magnitude larger still, with physical data-center space and contracted compute becoming the binding constraints rather than capital alone.

 

Nvidia builds its own open models under the Nemotron line and has been expanding its software and services footprint alongside its dominant position in AI accelerators. Access to Poolside’s Model Factory gives the chipmaker a specialized platform for producing generative AI systems tailored to software development. The non-exclusive nature of the license leaves Poolside free to offer the same technology to other customers.

 

The 109 employees receiving Nvidia job offers represent a substantial portion of Poolside’s technical workforce. Poolside’s chief executive has previously stated that fewer than 70 people built the company’s core model and that the combined engineering and research headcount was under 115. The talent transfer therefore moves a concentrated group of model-building specialists into Nvidia while the founding team continues to lead the independent company.

 

Industry observers note that such licensing-plus-hiring structures allow large technology companies to secure intellectual property and people without triggering the full regulatory scrutiny that often accompanies outright acquisitions. Similar arrangements have become more common as competition for AI talent and specialized model-training systems intensifies.

 

Poolside has spun out an infrastructure business that is developing a 1.2-gigawatt data center in Texas. Leadership appointments at that entity predated the Nvidia announcement. The investor letter also sketched a longer-term thesis: human-level AI capabilities will eventually be commoditized by open-source models, while the greatest remaining value will lie in problems that require real-world experimental feedback loops rather than pure intelligence.

 

The company suggested future focus areas could include scientific discovery domains such as drug development and materials research.

 

Nvidia has not issued a public statement on the specific valuation or strategic rationale for the $6 billion licensing fee. Reporting from The Information noted that the size of the payment relative to the technology being licensed remains opaque. 

 

Across its recent licensing and talent deals, Nvidia has committed tens of billions of dollars while keeping the target companies operational.

 

The transaction arrives as the AI industry continues to grapple with the escalating cost of frontier model training, the scarcity of advanced compute, and the strategic importance of specialized software stacks that sit above the hardware layer. 

 

For Nvidia, the deal deepens its vertical integration into the model-building process itself. For Poolside, it provides a large capital distribution to investors and a path to continue as an independent entity focused on whatever vision its founders choose to pursue next.

 

Market reaction and further details from both companies will determine how the arrangement shapes competition in open-weight coding models and enterprise AI tooling over the coming months. 

 

The structure itself, however, already illustrates one of the dominant deal patterns of the current AI capital cycle: large licensing payments, selective talent moves, and the preservation of independent operating companies rather than full absorption.