Nvidia is effectively centralizing the artificial intelligence supply chain by bypassing traditional mergers and acquisitions in favor of a repeatable, non-traditional financial mechanism. By deploying a licensing-plus-hire strategy, the company absorbs the operational capacity of critical AI startups while avoiding the regulatory scrutiny that typically accompanies full-scale corporate integration. This approach, most recently evidenced by the $7 billion deal with Poolside AI, allows Nvidia to secure the underlying infrastructure of innovation while leaving the target entities technically independent.
The transaction, first reported by Newcomer on August 20 based on a Poolside investor letter and subsequently confirmed by Bloomberg and The Information, involves a $6 billion payment for a non-exclusive license to Poolside’s Model Factory — the system the company uses to build its open-weight Laguna coding model family. In structural terms, a non-exclusive license grants Nvidia the right to utilize the technology while allowing Poolside to retain ownership and continue licensing it to other parties. Simultaneously, Nvidia is investing $1 billion in the company at a $12 billion pre-money valuation, a figure that quadruples Poolside’s previous $3 billion valuation. As part of the arrangement, 109 Poolside employees are transferring to Nvidia, though the company’s three co-founders, including former GitHub CTO Jason Warner and Eiso Kant, remain to lead the independent entity. The $6 billion licensing fee is expected to be distributed to existing backers — including Bain Capital Ventures, eBay, Citi Ventures, Redpoint, and Adams Street — by the end of 2027.
This is the third time Nvidia has run the same play. In December 2025, Nvidia paid approximately $20 billion for a non-exclusive license to Groq’s inference technology, with Groq founder Jonathan Ross and other staff joining Nvidia while the company continued independently under a new CEO. Earlier in 2025, Nvidia executed a similar arrangement with Enfabrica for AI networking hardware at roughly $900 million. By structuring each deal as a licensing agreement and minority equity investment rather than a buyout, Nvidia avoids the antitrust scrutiny that would accompany absorbing these companies entirely. The founders remain in place, and the companies continue to operate — providing the appearance of market competition while Nvidia’s control over the technical talent and proprietary systems deepens.
The focus on the Model Factory reveals what Nvidia is actually buying: not the output, but the mechanism of production. Poolside’s Laguna coding models — the Laguna M.1 scored 72.5 percent on SWE-bench Verified, competitive with Qwen-3.5 and DeepSeek V4-Flash but trailing Claude Sonnet 4.6 — are solid but do not currently lead the frontier. The smaller Laguna XS 2.1 variant hit 70.9 percent on SWE-bench Verified, and the long-horizon Laguna S 2.1 reached 70.2 percent on Terminal-Bench 2.1. These are credible coding models, but Nvidia is not paying $7 billion for benchmark rankings. It is paying for the industrial capacity to build coding models at scale. By securing the factory, Nvidia ensures it remains the primary beneficiary of the Laguna family’s development, regardless of which specific model version eventually captures the most market share.
The deal also functions as a structural bypass of standard regulatory oversight. Because it is framed as a licensing agreement and a minority equity stake rather than an acquisition, it sidesteps the scrutiny applied to horizontal or vertical integration. The 109 employees transferring to Nvidia represent a selective, high-value talent acquisition — enough to bolster Nvidia’s internal capabilities without the complexity of integrating an entire corporate structure. For Poolside’s investors, the $6 billion licensing payout by 2027 transforms what might have been a modest venture return into an extraordinary one, creating a powerful incentive structure that rewards this model of consolidation.
The Poolside deal extends the compute landlord thesis into new territory. Nvidia’s strategy now spans silicon-level investments in companies like Etched and Lancium, software-level control via the Model Factory, and deployment-level commitments to entities like OpenAI and the $500 billion Wall Street consortium. The SSI investment in July added another node to this network. By controlling the compute, the networking, and now the model-building factories, Nvidia is ensuring that every layer of the AI value chain is tethered to its own operational requirements.
If Nvidia continues to use its cash reserves to license the core systems of promising startups while absorbing their best talent, the industry may develop a hollowed-out version of independence. Poolside will continue to operate, but its most valuable asset — the ability to build models — is now effectively an extension of Nvidia’s internal R&D. As this playbook repeats, the distinction between an independent AI startup and an Nvidia-controlled subsidiary becomes increasingly difficult to discern. The market retains the appearance of diversity. The underlying technical reality is increasingly centralized around a single company that never had to file a single acquisition notice.