Nvidia licenses Groq inference tech, hires founders

Axios reports that Nvidia and Groq have signed a non‑exclusive inference technology licensing agreement aimed at making large language model inference cheaper and faster, with Groq’s language processing unit chips focused on real‑time chatbot responses. The deal, reached in late December 2025, will see Groq founder Jonathan Ross, president Sunny Madra and other staff join Nvidia while Groq continues to operate independently, and analysts say the structure resembles an acquihire designed to bolster Nvidia’s hold on the AI market’s costly inference phase while minimizing antitrust scrutiny.