Industry·3 min read
By BitsMindsSource: The Next Web

Nvidia Pays Poolside $6B to License Its Model Factory

Nvidia is licensing Poolside’s model-building software for $6 billion, investing another $1 billion at a $12 billion valuation, and hiring the 109 engineers who built its open-source coding model. Poolside insists it is not an acquisition.

NON-EXCLUSIVE LICENCE $6B for Model Factory · 109 engineers hired BITSMINDS.COM
Share:

Nvidia has agreed to pay Poolside $6 billion for a non-exclusive licence to Model Factory, the system the startup built to train AI models, and will make job offers to 109 of its employees. Separately, Nvidia is investing $1 billion in what remains of the company at a $12 billion pre-money valuation. Newcomer reported the terms first, drawing on an investor letter.

Poolside is adamant about what this is not. The letter states the deal is “not an acquisition and it is not an acquihire.” All three co-founders stay, the company keeps operating independently, and because the licence is non-exclusive it remains free to sell the same technology to someone else. The $6 billion is earmarked for distribution to investors by the end of next year.

The 109 engineers heading to Nvidia are the group that built Laguna, Poolside’s open-source coding model, which the company positioned as a Western counterweight to the open-weight releases coming out of DeepSeek and Qwen. Chief executive Eiso Kant has been pointed about how small that team was: “Less than 70 people built this model,” with fewer than 115 across engineering and research combined. That works out to roughly $55 million per hire — though the licence, not the headcount, is what the $6 billion nominally buys.

What pushed Poolside here was compute. The company missed a $2 billion funding window in January that would have paid for a 40,000-chip cluster, and without that hardware it lost the ability to train at the frontier. Model Factory is the asset that survived: not a model, but the machinery for building them — and Nvidia, which sells the chips such machinery runs on, is a natural home for a training stack that drags GPU demand along behind it.

This is the third deal Nvidia has structured this way. It paid roughly $20 billion in December 2024 for a non-exclusive licence to Groq’s inference designs and took on most of its staff and founders, leaving a rump company later valued at $3.5 billion. Enfabrica, a networking-hardware startup, went for about $900 million on similar terms. Across the three, Nvidia has committed something near $27 billion without filing a single merger notification.

That is rather the point of the structure. A licence-plus-hiring arrangement leaves a functioning independent company behind, which makes it harder for antitrust regulators to characterise the transaction as consolidation — even when the acquirer walks away with both the technology and the people who built it. Whether the distinction holds up under scrutiny at this scale is untested; what is clear is that it has now been used often enough to count as a playbook rather than an improvisation.

Investors were not uniformly impressed. Nvidia shares closed the week about 5 percent lower, a reminder that a company spending tens of billions to absorb its own ecosystem invites questions about how much of that ecosystem was ever going to stand up on its own. It also sits alongside Nvidia’s $105 billion guarantee for an OpenAI data center and a $63 billion equity portfolio weighted toward its own customers — three different mechanisms, all pointing the same direction.

Want AI news before everyone else?

The morning's most important AI stories, straight to your inbox. No fluff.

Related Articles

Editorial illustration: a monitor showing the ring of twelve European stars, one knocked out of place and marked by a red target reticle, beside the word ARTEX. A metaphor for the reported AI-assisted cyberattack on the European Commission; not an official EU image. ARTEX AI-ASSISTED ATTACK BITSMINDS.COM
Industry

European Commission Hit by AI-Assisted ARTEX Attack

California sues OpenAI over its rogue agents Editorial illustration on a charcoal field: a polished steel gavel striking a round steel sound block whose top face carries the OpenAI mark as an engraving. Left, the words The People of California versus OpenAI and Rogue agents. An editorial metaphor for a lawsuit filed in San Francisco Superior Court, not a depiction of any court proceeding. THE PEOPLE OF CALIFORNIA v. OpenAI ROGUE AGENTS BITSMINDS.COM
Industry

California Sues OpenAI Over Its Rogue AI Agents

US rule on remote access to advanced AI chips Editorial illustration on an oxblood field: an angled AI accelerator package with a metal lid and four memory stacks, with a thick network cable plugged into its side and a brass padlock engraved LICENSE locked around the cable. The type column reads U.S. export controls, Remote access, license required. The padlock is an editorial metaphor for the Commerce Department's new licence requirement on overseas data centres; the chip is generic and is not a specific vendor's product. 3A090 3A090 ADVANCED IC LICENSE U.S. EXPORT CONTROLS Remote access LICENSE REQUIRED BITSMINDS.COM
Industry

US Closes AI Chip Remote-Access Loophole With Licenses