Chip giant licenses Groq's inference technology and hires top executives, joining Big Tech's acquihire spree to dominate AI's next frontier.
NVIDIA has agreed to license artificial intelligence inference chip technology from Groq under a set of commercial licensing and employment agreements announced Dec. 25, 2025, securing rights to use Groq-developed technology and bringing senior Groq personnel into Nvidia while leaving the startup operating independently.
The agreement involves Nvidia and AI chip startup Groq, and is structured around a non-exclusive technology license paired with targeted hiring rather than a full corporate acquisition. Public reporting has widely placed the overall transaction value at roughly $20 billion, though the arrangement was not described as a purchase of the company.
The immediate effect of the deal is that Nvidia gains licensed access to Groq’s inference-focused chip technology alongside selected engineering assets and senior executives, while Groq continues to exist as a standalone entity under a revised leadership structure.
The Inference Battleground
Groq specializes in what is known as inference, where artificial intelligence models that have already been trained respond to requests from users. While Nvidia dominates the market for training AI models, it faces much more competition in inference, where traditional rivals such as Advanced Micro Devices have aimed to challenge it, as well as startups such as Groq and Cerebras Systems.
Groq has been working on a different type of chip called an LPU (language processing unit), which the company claims can run large language models ten times faster while using one-tenth the energy of conventional chips.
Procedural structure of the licensing and hiring agreement
The core of the transaction is a non-exclusive license granting Nvidia the right to use Groq’s AI inference chip technology. The license allows Nvidia to incorporate the technology into its own product development efforts without transferring ownership of Groq as a corporate entity. In addition to the technology license, the agreement provides for the transfer of selected assets and engineering talent from Groq to Nvidia.
These transfers are tied to employment arrangements rather than equity transactions, placing personnel moves alongside the licensing terms as a central component of the deal. The structure explicitly avoids a full acquisition of Groq. No merger agreement was announced, and the transaction was not presented as a takeover of the startup or a consolidation of corporate control.
From a legal and regulatory standpoint, the use of licensing and hiring agreements positions the deal outside the framework of an outright merger. By relying on commercial licensing contracts and employment arrangements, the companies avoided the procedural requirements associated with acquiring an entire operating company.
Positions taken by Nvidia and Groq
Nvidia described the arrangement as an expansion of its technology portfolio and a means of acquiring specialized talent in AI inference. Public descriptions focused on access to Groq’s technology and the addition of experienced personnel rather than the purchase of the startup itself.
Groq, for its part, confirmed that it will continue operations as a standalone company. The deal does not dissolve Groq’s corporate existence, and its remaining business activities are expected to proceed independently following the personnel changes.
While the transaction has been widely reported at a value of approximately $20 billion, neither company characterized the figure as a purchase price for Groq. The reported valuation reflects the scale attributed to the combined licensing rights and talent agreements rather than an equity transfer.
Integration and next steps following the announcement
Under the terms outlined, Groq is expected to operate under new leadership following the departure of selected executives and engineers who are joining Nvidia. The company remains independent and continues to exist as a separate operating entity. Nvidia plans to integrate the licensed AI inference technology into its product offerings. The agreement provides Nvidia with rights to use the technology internally, aligning it with the company’s existing development and deployment processes.
No timeline was disclosed for the completion of technology integration or for the full onboarding of Groq personnel. The announcement focused on the structure and scope of the agreement rather than implementation schedules or future performance targets.
The companies did not announce additional transactions or follow-on agreements as part of the Dec. 25 disclosure. The deal, as described, consists of the non-exclusive license, the transfer of selected assets, and the hiring of senior Groq personnel, with Groq continuing operations independently.
