Groq raises $650M to rebuild AI inference cloud after Nvidia's $20B licensing deal stripped key talent

AI Chipmaker Groq Secures $650M Funding After Nvidia Deal

AI chipmaker Groq has confirmed a $650M funding round led by growth investment firm Disruptive and hedge fund Infinitum, marking a dramatic comeback roughly six months after Nvidia struck a non-exclusive licensing agreement for Groq's technology 1

1

. The deal, reported to be worth around $20 billion, saw Nvidia acquire the intellectual property for Groq's Language Processing Unit (LPU) chips while simultaneously hiring away founder and CEO Jonathan Ross, president Sunny Madra, and other critical engineers 2

2

. Groq did not disclose its new valuation, though it was last valued at $6.9 billion following a $750 million round in September 1

1

.

Pivoting to Groq's Neocloud Business Model

With Nvidia now controlling the LPU intellectual property and launching its own Nvidia Groq 3 LPX inference hardware system at its GTC event in March, Groq has pivoted hard toward its neocloud business 1

1

. The company now operates 13 data centers across North America, Europe, the Middle East, and Asia-Pacific, serving over five million developers and thousands of AI companies while processing trillions of tokens each week 3

3

. The new capital will be deployed to fit out this footprint with latest hardware, including ironically the Nvidia LPX system built on licensed Groq designs, with plans to quadruple total capacity to 200 megawatts by the end of 2027 2

2

.

New Leadership Takes the Helm

Following Jonathan Ross's departure to Nvidia, Groq has assembled an entirely new executive team. Doug Wightman, the Google engineer who co-founded Groq with Ross a decade ago, stayed on after the Nvidia deal and became CEO 1

1

. The company added Alan Rice as COO, previously at xAI and Meta, after a career in the U.S. Navy 1

1

. From July, Sinclair Schuller joins as CTO and Rakesh Malhotra as CPO, an entrepreneurial duo who previously built and sold enterprise software firms Apprenda and Nuvalence, with Malhotra spending roughly a decade on Microsoft's cloud products 2

2

.

The AI Infrastructure Race Intensifies

Groq is wading into a brutal AI compute market where demand for AI inference is climbing fast and money is pouring in around it 2

2

. AI-infrastructure startup Baseten recently raised $1.5 billion at a valuation of up to $13 billion, while frontier labs OpenAI and Anthropic keep pushing the cost of compute higher, which only sharpens the appetite for cheaper ways to run their models 2

2

. Other companies like CoreWeave have also broadened their focus beyond infrastructure to higher-level services 3

3

. Groq's pitch centers on the argument that as AI shifts from training models to running them, purpose-built inference chips should beat general-purpose GPUs on speed and cost, with inference eventually needing 15 to 20 times more compute than training 2

2

.

Can Groq Compete Against Its Former IP Owner?

Whether Groq can succeed after essentially selling its core technology depends on how competitive its inference cloud service can remain now that the key hardware IP is shared with Nvidia 1

1

. The LPU 3 includes advanced features like automatic clock drift correction to avoid data traffic bottlenecks, 92 lanes moving data at 112 gigabits per second for 2.5 terabits per second of bidirectional bandwidth, and 500 megabytes of onboard SRAM for faster inference 3

3

. In an unusual twist, many of the backers now writing checks are the same ones Nvidia cashed out in December, with Disruptive and Infinitum both holding board seats 2

2

. Precedent suggests survival is possible: Scale AI's CEO told Forbes that business rebounded after Meta did a $14.3 billion not-acqui-hire about a year ago, and that the company is on track to do $1 billion in revenue 1

1

. One way Groq could set itself apart is by extending its AI-optimized cloud platform with new services such as managed databases, following the playbook of other AI-focused cloud providers 3

3

.