Nvidia has agreed to a major arrangement with AI startup Poolside that combines a $6 billion non-exclusive license for the company’s Model Factory platform with a $1 billion equity investment valuing Poolside at $12 billion pre-money. 

As part of the deal, Nvidia is extending job offers to 109 Poolside employees while the startup’s three co-founders remain in place and the company continues to operate independently.

The terms come from a letter Poolside sent to its investors, first reported by Newcomer and subsequently confirmed by Bloomberg, The Information, and other outlets in the days surrounding August 20–21, 2026.

Poolside has described the transaction as “not an acquisition and it is not an acquihire.” The $6 billion licensing fee is expected to be distributed to existing investors by the end of 2027.

What Nvidia Is Licensing

The centerpiece of the agreement is a non-exclusive license to Poolside’s Model Factory, the internal system the startup has used to develop its Laguna family of open-weight coding models. 

These models are designed for agentic software engineering tasks: planning, editing code, running tests, interpreting results, and iterating over extended periods.

Poolside has released several Laguna variants in 2026, including smaller and larger mixture-of-experts configurations that emphasize coding performance and long-horizon autonomy. 

The models have been trained and optimized to run efficiently on Nvidia hardware, including consumer- and enterprise-grade systems. By licensing the Model Factory, Nvidia gains access to the tooling, workflows, and infrastructure that produced these models without taking full ownership of the company.

Because the license is non-exclusive, Poolside retains the ability to continue developing and licensing its technology to other parties. The arrangement therefore expands Nvidia’s internal capabilities while leaving Poolside free to pursue its own roadmap.

Equity Investment and Talent Component

Alongside the license, Nvidia is investing $1 billion in Poolside at a $12 billion pre-money valuation. The investment gives Nvidia a direct financial stake in the company’s future performance and aligns the two firms more closely on technical collaboration.

Separately, Nvidia has made job offers to 109 Poolside employees who worked on the Laguna models and related systems. The three co-founders, including former GitHub CTO Jason Warner, are remaining with Poolside to lead the independent company.

This structure—license plus targeted hiring plus minority equity—mirrors previous Nvidia deals with other specialized AI firms and has drawn attention for allowing the chipmaker to secure technology and talent without triggering full acquisition review processes.

Why the Deal Matters

The transaction underscores several broader trends in the AI industry. First, the value of specialized model-building infrastructure and high-performing coding agents has risen sharply. 

Coding models and the systems that produce them sit at the intersection of enterprise demand, open-weight competition, and the race to improve agent reliability on multi-step software tasks.

Second, Nvidia continues to deepen its involvement beyond hardware. While the company remains the dominant supplier of AI accelerators, it has increasingly invested in software platforms, open models (such as its Nemotron line), and strategic relationships with model developers. 

Access to Poolside’s Model Factory strengthens Nvidia’s ability to iterate on coding-focused systems that can run on its own chips.

Third, the deal structure itself has become a recurring pattern. By packaging a large license fee, selective hiring, and an equity stake, Nvidia can accelerate internal capabilities while the target company continues as an independent entity. 

Similar approaches have been used with other firms in the AI ecosystem. Critics have noted that such arrangements can achieve many of the practical effects of an acquisition while remaining outside traditional merger scrutiny; supporters argue they enable faster technology transfer without the overhead of full integration.

Poolside’s Trajectory

Poolside was founded in 2023 and has focused on coding agents and open-weight models optimized for software development. 

Earlier funding included a $500 million Series B in 2024 that already counted Nvidia among its investors. The new valuation of $12 billion pre-money represents a substantial step-up and reflects the market’s appetite for specialized coding and agentic systems.

The Laguna models have been positioned as competitive open-weight alternatives in the coding domain, with strong results on agentic benchmarks and the ability to run extended autonomous workflows. 

Poolside has also released supporting tools, including a desktop application and command-line harness intended to make multi-agent coding sessions more practical for developers.

The $6 billion licensing distribution planned for existing backers provides a significant liquidity event while the company itself continues operating with fresh capital and retained leadership.

Implications for the Broader Market

For developers and enterprises, the deal may accelerate the availability of higher-quality coding agents and the underlying tooling used to train them. 

Nvidia’s involvement increases the likelihood that Model Factory-derived capabilities will be optimized for its hardware ecosystem, potentially improving performance and efficiency for users already standardized on Nvidia GPUs.

For the competitive landscape, the transaction highlights the strategic importance of the “factory” layer—the software, data pipelines, evaluation harnesses, and iteration loops that turn raw compute into capable models. 

As frontier labs and open-weight developers race to improve agent reliability, control over these internal systems is becoming as valuable as the models themselves.

The arrangement also illustrates the continuing consolidation of talent and expertise around a small number of well-capitalized platforms. 

While Poolside remains independent, the transfer of 109 employees and the licensing of core technology effectively integrate a significant portion of its technical capacity into Nvidia’s broader AI efforts.

What Comes Next

Neither company has publicly detailed the exact timeline for integrating the licensed technology or the specific product outcomes expected from the collaboration. Poolside is expected to continue releasing and improving Laguna models under its own brand while distributing the licensing proceeds to investors. 

Nvidia is likely to incorporate Model Factory capabilities into its internal model development and possibly into future open or commercial offerings.

Observers will watch whether similar licensing-plus-equity structures become more common as large technology companies seek specialized AI capabilities without full acquisitions. 

They will also monitor how quickly the licensed technology translates into measurable improvements in coding agents, open models, or developer tools that reach the market.

The $7 billion commitment ranks among the larger recent transactions in the AI sector and reinforces Nvidia’s dual strategy of supplying the hardware that powers the industry while securing preferential access to the software systems that turn that hardware into capable AI products.