Nvidia Seals a Landmark $20 Billion Deal for Groq’s AI Chip Tech
In a record-breaking deal, Nvidia has committed to acquiring assets from AI chip startup Groq for an estimated $20 billion in cash. This strategic move aims to reinforce Nvidia’s foothold in the rapidly expanding AI inference market.
The transaction, framed as a non-exclusive licensing agreement rather than a typical acquisition, will enable Nvidia to tap into Groq’s Language Processing Unit (LPU) technology. Additionally, key personnel, including founder and CEO Jonathan Ross—who played a pivotal role in the development of Google’s Tensor Processing Unit—will join Nvidia. The plan is to incorporate Groq’s low-latency processors into Nvidia’s AI factory architecture to cater to a wider range of AI inference and real-time workloads.
Despite Nvidia’s dominance in the AI training chip market, the company is facing growing competition in the inference space. This is where AI models generate responses to user queries. Groq’s specialized chips, which utilize on-chip SRAM memory technology, enable incredibly fast per-token processing. This makes them especially suitable for hosting chatbots and real-time AI applications. The deal is Nvidia’s largest to date, dwarfing its 2019 acquisition of Israeli chip designer Mellanox, which was close to $7 billion.
Groq will maintain its independence, with Simon Edwards at the helm as CEO. The company’s GroqCloud business will continue to operate without any disruption. Just three months ago, in a funding round in September, the startup was valued at $6.9 billion.
Source: CNBC