Groq raises $650M, pivots to neocloud inference after Nvidia's $20B license deal
Groq announced a $650 million funding round on June 22, 2026, led by Disruptive and Infinitum, just six months after Nvidia paid roughly $20 billion in December 2025 for a non-exclusive license to Groq's chip technology and hired away founder-CEO Jonathan Ross and president Sunny Madra. The round validates investor confidence that a hardware startup can survive a major talent exodus by pivoting to services—Groq is now betting on its neocloud inference business rather than selling chips directly.
The AI chipmaker operates 13 data centers across North America, Europe, the Middle East, and Asia-Pacific, serving over 5 million developers and processing trillions of tokens weekly. Groq plans to expand that infrastructure dramatically, targeting 200 megawatts of capacity by end of 2027 and deploying Nvidia's new LPX platform (which incorporates Groq's licensed inference tech) alongside Groq's own LPU silicon. The company has also rebuilt leadership, adding Alan Rice (formerly xAI, Meta) as COO, with CTO Sinclair Schuller and Chief Product Officer Rakesh Malhotra joining in July.
The financing sends a clear message: infrastructure matters more than chip designers. Groq's recovery signals that even after a competitor acquires your talent, a defensible platform—global datacenters, proven developer base, neocloud positioning—can sustain venture funding. For architects: inference commoditization is real, and it's accelerating who survives in the chip business.