Under the agreement, Nvidia will obtain a non-exclusive license to Groq’s chip technology. At the same time, Groq founder Jonathan Ross—who previously played a key role in launching Google’s AI chip initiative—along with Groq President Sunny Madra and several senior engineers, will join Nvidia. A source close to Nvidia confirmed the licensing arrangement, though financial details were not disclosed.
Groq emphasized that it will continue operating as an independent company, with Simon Edwards stepping in as CEO. Its cloud services business will also remain active. This clarification followed media reports suggesting Nvidia had agreed to acquire Groq for $20 billion in cash—claims neither company commented on publicly.
Groq specializes in AI inference, the stage where trained artificial intelligence models generate responses to user queries. While Nvidia has established near-dominance in AI training hardware, inference has become a more competitive arena. Rivals include established chipmakers such as AMD, as well as startups like Groq and Cerebras Systems, all aiming to capture market share as AI applications scale.
The structure of the deal mirrors several recent high-profile arrangements in the tech industry. Microsoft, Meta, Amazon, and Nvidia itself have all pursued similar strategies—paying large sums for licenses and talent while avoiding formal acquisitions. These arrangements have drawn increasing regulatory attention, though none have been reversed so far.
Commenting on the deal, Bernstein analyst Stacy Rasgon noted that antitrust scrutiny remains the primary risk. He suggested that framing the agreement as a non-exclusive license may help preserve the appearance of competition, even as Groq’s leadership and technical expertise shift to Nvidia. Rasgon also pointed out that Nvidia CEO Jensen Huang is widely viewed as having a strong relationship with the current U.S. administration, potentially reducing regulatory friction.
Groq’s profile has risen sharply over the past year. Following a $750 million funding round in September, the company’s valuation more than doubled to $6.9 billion, up from $2.8 billion the previous August. The company is part of a group of AI chip startups that avoid reliance on external high-bandwidth memory, instead using on-chip SRAM. This design choice reduces exposure to global memory shortages and enables faster response times for chatbots and similar applications, though it also limits the size of models that can be deployed.
Cerebras Systems remains Groq’s closest competitor in this approach and is reportedly preparing for a potential public offering as early as next year. Both companies have also secured significant contracts in the Middle East, highlighting growing global demand for inference-focused AI infrastructure.
Nvidia CEO Jensen Huang has repeatedly stressed that the company is well positioned to maintain its leadership as AI workloads shift from training to inference. During his largest keynote address of 2025, Huang argued that Nvidia’s ecosystem and strategic investments would allow it to stay ahead even as competition intensifies.
Overall, the Groq deal underscores Nvidia’s intent to strengthen its position in the inference market through strategic licensing and talent acquisition, while avoiding the regulatory hurdles associated with full-scale mergers and acquisitions.