Groq’s AI Cloud Faces a New Test After Nvidia Deal

Nvidia paid roughly $20 billion last December to license Groq’s signature chip architecture and hire its leadership, stopping short of a full acquisition, according to Business Insider. The deal strengthened Nvidia’s position in AI inference, the market for running AI models. For Groq, it left behind a company with real infrastructure but without the core chip advantage that once made it stand out as an Nvidia challenger.

Groq remains a sizable AI cloud business, with 13 data centers, plans to quadruple capacity next year, and 6 million developers plus thousands of companies using its cloud. Business Insider reports that Groq is backed by $1 billion in fresh capital and was valued at $3.5 billion in its latest funding round, down from $6.9 billion before the Nvidia deal, though shareholders received payouts in between. The company is now betting that demand for AI compute is large enough to support multiple winners.

Some customers told Business Insider they experienced weaker technical support or reduced AI model choice after the deal. Recall CEO Paul Richards said uncertainty led his company to rely less on Groq, while Willow cofounder Lawrence Liu described a short “re-initiation period” after key support staff moved to Nvidia. Others saw fewer problems: Sourcetable founder Eoin McMillan said the experience was “business as usual.”

Groq is now competing in an AI inference market where providers are proliferating and speed advantages are narrowing. Business Insider names rivals including Fireworks, Together AI, CoreWeave, Amazon, and Google. The key question is whether Groq can win customers on economics, capacity, reliability, or its experience running language processing units at scale.