Nvidia has agreed to license chip technology from startup Groq and hire the veteran CEO of Alphabet Inc.'s Google, Groq said in a blog post Wednesday.
Groq specializes in so-called inference, where already trained artificial intelligence models respond to requests from users.
Nvidia dominates the market for training AI models, but it faces even more competition in inference, with traditional rivals like Advanced Micro Devices and startups like Groq and Cerebras Systems looking to challenge it.
Groq said Nvidia has agreed to a “non-exclusive” license to Groq's technology. Groq said founder Jonathan Ross, who helped launch Google's AI chip program, as well as Groq president Sunny Madra and other members of the engineering team, will join NVIDIA.
A person close to Nvidia confirmed the licensing agreement.
Groq did not disclose financial details of the transaction. CNBC previously reported that Nvidia had agreed to buy Groq for $20 billion in cash, but neither Nvidia nor Groq would comment on the report. Groq said in a blog post that it will continue to operate as an independent company with Simon Edwards as CEO and will continue its cloud business.
Groq more than doubled its valuation to $6.9 billion from $2.8 billion last August following a $750 million funding round in September.
Groq is one of the startups that does not use external high-bandwidth memory chips, thereby relieving itself of the memory shortage affecting the global chip industry. This approach, which uses a form of on-chip memory called SRAM, can help speed up interactions with chatbots and other AI models, but it also limits the size of models that can be delivered.
Groq's main rival in this approach is Cerebra Systems, which Reuters reported earlier this month plans to go public as early as next year. Grok and Cerebras both signed major contracts in the Middle East.
Nvidia CEO Jensen Huang spent much of his biggest keynote of 2025 arguing that Nvidia can maintain the lead as the AI market shifts from training to inference.
