Nvidia Licenses Groq Inference Technology, Hires Key Executives
Nvidia has agreed to license chip technology from startup Groq and to hire its top executives, according to a company blog post released Wednesday. The arrangement fits a broader pattern. In recent years, major technology firms have paid heavily for access to startup technology and talent. However, they have often avoided outright acquisitions.
Under the agreement, Nvidia will license Groq’s chip technology for inference workloads. At the same time, Groq founder Jonathan Ross, Groq President Sunny Madra, and several engineers will move to Nvidia. A person familiar with the matter confirmed the licensing deal. Still, financial terms were not disclosed.
Focus on inference competition
Groq specializes in inference, the stage where trained artificial intelligence models generate responses for users. While Nvidia leads the market for training AI models, it faces stronger competition in inference. As a result, rivals such as AMD and several startups have pushed aggressively into this space.
Nvidia agreed to a non-exclusive license, Groq said. Meanwhile, Groq stated it will continue operating independently, with Simon Edwards serving as chief executive. In addition, its cloud business will remain active. CNBC reported that Nvidia had agreed to acquire Groq for $20 billion in cash, but neither company commented on the report.
Regulatory scrutiny and market context
Similar arrangements have drawn attention across the industry. For example, Microsoft brought in its top AI executive through a $650 million licensing-style deal. Likewise, Meta spent $15 billion to hire the CEO of Scale AI without buying the firm. Amazon and Nvidia have pursued comparable talent-focused deals. Consequently, regulators have begun to scrutinize these transactions, although none has been reversed so far.
“Antitrust would seem to be the primary risk here, though structuring the deal as a non-exclusive license may keep the fiction of competition alive (even as Groq’s leadership and, we would presume, technical talent move over to Nvidia),” Bernstein analyst Stacy Rasgon wrote in a note to clients on Wednesday after Groq’s announcement. And Nvidia CEO Jensen Huang’s “relationship with the Trump administration appears among the strongest of the key US tech companies.”
Groq more than doubled its valuation to $6.9 billion after a $750 million funding round in September, up from $2.8 billion last August. The company uses on-chip SRAM memory instead of external high-bandwidth memory. As a result, it avoids industry-wide memory shortages and speeds up chatbot responses, although model size is limited. Its main rival, Cerebras Systems, is planning an initial public offering as soon as next year. Meanwhile, Nvidia’s Huang has argued publicly that the company can maintain its lead as AI markets continue shifting from training toward inference.
