Groq confirms $650M raise and restaffs after Nvidia deal

Groq announced a new $650 million funding round on Monday, roughly six months after Nvidia signed a non-exclusive licensing agreement in December and hired away founder and CEO Jonathan Ross, president Sunny Madra, and other employees. The startup did not disclose a new valuation; it was last valued at $6.9 billion after a $750 million round in September.

How did Groq get here?

Groq launched about a decade ago when Jonathan Ross and Doug Wightman left Google to build an AI chip company; Ross later departed after Nvidia’s licensing deal in December. Ross had been known for helping create Google’s Tensor Processing Unit, and he teamed with Wightman to found Groq. After the Nvidia agreement, Wightman stayed and became CEO while Groq’s core LPU hardware IP became available to Nvidia.

Groq designed a language processing unit, or LPU, intended for inference and sold it as part of a cloud service or on-premises hardware cluster. With Nvidia now owning the IP for LPUs, Nvidia introduced its own Nvidia Groq 3 LPX inference hardware system at GTC in March.

What is Groq doing now?

Groq is pivoting to a neocloud business and rebuilding its executive team while expanding its global data center footprint. The neocloud unit, which Groq says was run by Sunny Madra after Groq acquired his company Definitive Intelligence in 2024, has grown to 13 data centers across North America, Europe, the Middle East, and APAC. Groq says the neocloud serves over five million developers and thousands of AI companies, processing trillions of tokens each week.

To replace departed leaders, Groq hired Alan Rice as chief operating officer; Rice previously worked at xAI and Meta and served in the U.S. Navy. Groq also added Sinclair Schuller as chief technology officer and Rakesh Malhotra as chief product officer. Schuller and Malhotra previously worked together at Apprenda, then co-founded Nuvalence, a software-engineering firm acquired by EY in 2024. Malhotra spent about a decade working on Microsoft’s cloud products.

Why it matters

Groq’s situation illustrates how IP licensing and talent movement can reshape an AI hardware vendor in months. Nvidia’s licensing of Groq’s LPU IP and its March launch of the Groq 3 LPX cluster put Groq in direct competition with a company that now sells similar inference hardware. At the same time, demand for inference infrastructure remains strong, and Groq’s neocloud already claims scale: 13 data centers and more than five million developers is a nontrivial footprint.

The company’s ability to convert that footprint into sustainable revenue will determine whether it follows other companies that rebounded after large not-acqui-hire deals. The article notes that Scale AI’s CEO Jason Droege said business rebounded after Meta’s $14.3 billion not-acqui-hire and that Scale is on track to do $1 billion in revenue, an example investors and customers will watch.

What to watch

Watch for a disclosed valuation following the $650 million raise; Groq declined to disclose one, and its last public valuation was $6.9 billion in September. Monitor customer growth and usage metrics from the neocloud, and any technical or pricing differentiation Groq announces in response to Nvidia’s Groq 3 LPX hardware. Executive stability and the company’s ability to retain and scale engineering talent will also be a concrete signal of whether the pivot can succeed.

Groq’s next milestones are straightforward: reveal how the $650 million will be allocated, report revenue or customer traction for the neocloud, and outline product differences that separate its inference service from Nvidia’s newly announced hardware cluster.