$20 billion to buy Groq—what does Nvidia want?

Nvidia caused a stir in Silicon Valley on Wednesday, agreeing to pay around $20 billion for a technology license from the startup Groq and to hire its core team. This huge deal is aimed not only at consolidating Nvidia's dominance in AI inference computing by acquiring Groq's specialized technology, but also uses a special transaction structure to circumvent increasingly strict antitrust scrutiny. According to a source involved in the deal, its specific form is a non-exclusive technology license, and Nvidia will also hire Groq’s founders and executives. Although full details of the deal have not yet been disclosed, the roughly $20 billion sum is nearly three times Groq’s $6.9 billion valuation during fundraising several months ago. With this move, Nvidia seeks **to obtain more cost-effective and faster chip design capabilities to meet the growing demand for running AI applications.** Nvidia CEO Jensen Huang clarified the strategic intent behind the deal in an internal email to employees. He said that the plan is to integrate Groq's low-latency processors into Nvidia's AI factory architecture, thus expanding platform capabilities to serve broader AI inference and real-time workloads. This means Nvidia is attempting to shore up its weakness in efficient inference chips, beyond its extremely expensive high-performance training chips. The structure of this deal mirrors the model adopted by Microsoft, Amazon, and Google over the past two years: bypassing formal company acquisitions — and regulatory scrutiny — using “licensing technology + hiring talent.” This move gives Nvidia not only key intellectual property and talent but also reflects how the world's most valuable company is **leveraging its $60 billion cash reserves to accelerate building defensive barriers** in the face of competition from rivals like Google’s TPU. **Shoring up the inference gap** The core driver of this transaction is Nvidia’s fight for the AI inference market. Although Nvidia’s GPUs and supporting software absolutely dominate AI model development and training, its existing chips are often too large and expensive to run practical applications (inference) like chatbots. **The market has long been searching for cheaper, more efficient alternatives, and Groq’s technology is designed for this purpose.** This license arrangement enables Nvidia to obtain Groq’s intellectual property. Groq claims its chips can process data faster than Nvidia’s when it comes to specific AI application tasks. By comparison, while Nvidia’s chips offer flexibility for multiple types of operations, there is room for optimization in speed and latency. Dylan Patel, chief analyst at chip consultancy SemiAnalysis, points out that while Groq's first-generation chips haven't yet formed strong competition with Nvidia, its next two generations are about to launch. He believes **Nvidia may have seen the threat in Groq’s new-generation technology, which is why it made its move.** **The “License + Talent Acquisition” Special Structure** This deal is not a traditional full acquisition. Groq founder Jonathan Ross, president Sunny Madra, and other employees will join Nvidia to “advance and expand” the licensed technology. At the same time, Groq’s original cloud business will remain internal, run by the newly appointed CEO Simon Edwards, who joined as CFO in September. This non-exclusive licensing structure is a common approach for tech giants seeking to avoid regulatory scrutiny. Microsoft, Amazon, and Google have all used similar structures to absorb founders and key tech from AI startups without a formal acquisition. Although Google’s similar transaction with Character.ai drew scrutiny from the U.S. Department of Justice, no action was taken. **Currently, Nvidia is not subject to U.S. antitrust review, but remains cautious in describing its market share in AI chips.** Sources told The Wall Street Journal that, as a result of the license agreement, Groq investors (including BlackRock and Tiger Global Management) will receive staged payments based on future performance. This deal is similar to Nvidia’s $900 million transaction three months ago with network startup Enfabrica, where Nvidia hired that company’s CEO and engineering team and paid a technology license fee. **The Unshakable Nvidia Ecosystem** Despite billions in venture funding, challengers like Groq have struggled to break Nvidia's tight control over the high-end AI chip market. The proprietary CUDA programming ecosystem of Nvidia’s chips creates high customer stickiness due to performance. Groq’s recent business situation also illustrates the difficulty of challenging giants. The company recently lowered its 2025 revenue forecast by about three-quarters. Groq’s spokesperson said this was because regions planned for chip deployment lacked data center capacity, causing some revenue expectations to be delayed until next year. Groq had earlier projected its cloud business would exceed $40 million in revenue this year, with total sales over $500 million. At the same time, competition is intensifying. Google’s TPUs are emerging as strong competitors to Nvidia GPUs, and major companies like Apple and Anthropic are using TPUs to train large models. Meta and OpenAI are also developing proprietary inference chips to reduce dependence on Nvidia. Among startups, consolidation is evident: Intel is in deep talks to acquire SambaNova, Meta acquired Rivos, and AMD absorbed the team from Untether AI. **Cash Reserves as an Economic Moat** Nvidia is using its large cash reserves to consolidate its business. As of the end of October, its cash reserves reached $60 billion. In addition to funding dozens of cloud providers and startups for exclusive purchases or rentals of its chips, Nvidia has begun large-scale technology acquisitions. Previously, Nvidia’s largest acquisition was the $6.9 billion purchase of Mellanox in 2019, which has become an important networking division contributing around $2 billion in last quarter’s revenue. Although the $20 billion deal with Groq is not a full acquisition, its scale far surpasses previous deals, showing that Nvidia is willing to pay a hefty price to eliminate potential threats and integrate cutting-edge technology in the face of increasingly specialized chip demands. Risk Warning and Disclaimer The market contains risks; investment requires caution. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situation, or needs. Users should consider whether any opinions, views, or conclusions in this article fit their specific circumstances. Investing based on this information is at one’s own risk.