Nvidia Pays $6 Billion to License Poolside's Model Factory and Offers Jobs to 109 Staff — and Both Sides Say It Is Not an Acquisition

Per an investor letter obtained by Newcomer, Nvidia struck a non-exclusive $6 billion licensing deal for Poolside's Model Factory — the infrastructure suite covering data processing, training, reinforcement learning and evaluation that Poolside uses to build its Laguna family of open-weight coding models — plus a $1 billion investment at a $12 billion pre-money valuation. Nvidia will extend offers to 109 employees who worked on the model; Poolside's three co-founders stay. Poolside plans to distribute the $6 billion to investors by the end of next year.

What Was Licensed Is the Assembly Line, Not a Model

The license does not cover a particular model but the ability to make them. Model Factory is the full infrastructure Poolside uses to produce its Laguna family of open-weight coding models, spanning data processing, training, reinforcement learning and evaluation. The license is non-exclusive, so Poolside can still sell the same technology to other buyers. Structurally the deal has three parts: a $6 billion license fee, a $1 billion growth round at a $12 billion pre-money ($13 billion post-money) valuation, and job offers to the 109 employees who worked on Laguna. The founders wrote to investors that "this is not an acquisition and it is not an acquihire" — worth quoting as-is, because it is the parties' own characterization rather than an outside conclusion. It does share a shape with the past year's deals: Scale and Meta, Character.AI and Google, Nvidia and Groq (about $20 billion), Nvidia and Enfabrica (about $900 million) — a large company takes the technology and the team and leaves the corporate shell behind. The difference here is the people. In earlier cases the founders went along; Poolside's three co-founders are staying. Nvidia was already an investor, having joined the company's $500 million Series B in October 2024.

Why Poolside Sold

The letter gives a concrete reason: continuing to compete in open-weight model development required more access to Nvidia hardware than the company could secure on its own. It also describes a six-week window at the end of last year to raise $2 billion for a 40,000-GB300 cluster coming online in January — funding they did not close in time, costing them the cluster. That account states the current barrier more bluntly than any industry analysis: at this level, open-weight model work is gated on compute procurement, not research ability.

Why This Structure Is Spreading

For the buyer, licensing plus hiring secures the technology and the people while avoiding the scrutiny and integration costs of a full acquisition. For the seller's investors, $6 billion distributed in cash beats waiting on an uncertain exit. The practical implication for readers: judging a model startup's position by its funding total is no longer enough — whether it can reliably secure accelerators at scale is becoming the dividing line. One caveat: this account comes from the investor letter obtained by Newcomer and follow-up reporting. The two companies have not issued a joint announcement, and final terms rest on eventual disclosure.

via: Newcomer, Bloomberg, The Next Web, The Decoder