Groq Funding: AI Chip Startup Raises $650 Million After Nvidia Talent Raid

by Zaira Varghese 23. June 2026 AI 0

Fresh Groq funding is back in the headlines after the AI chip startup confirmed a $650 million raise this week, money it will use to rebuild itself as an inference cloud provider following Nvidia’s $20 billion talent raid earlier in 2026. The round was led by existing backers Disruptive and Infinitum, both of which hold seats on Groq’s board.

It is a remarkable second act for a company that looked gutted just six months ago, when Nvidia signed a non-exclusive license for Groq’s technology and hired away founder and CEO Jonathan Ross, president Sunny Madra, and a chunk of the engineering team. Rather than fold, Groq pivoted—and investors clearly believe there is still a business worth backing.

What the new Groq funding round buys

This Groq funding will go toward expanding data center capacity and scaling the company’s inference cloud. Groq says its platform now processes trillions of tokens per week for around 5 million developers, running across 13 data centers spanning multiple continents. The stated goal is to reach 200 megawatts of inference capacity by 2027, a meaningful footprint in a market where compute is the bottleneck.

$650 million raised, led by Disruptive and Infinitum.
Proceeds aimed at expanding inference capacity to 200 MW by 2027.
Trillions of tokens processed weekly for roughly 5 million developers.
13 data centers already live across multiple continents.

How the Nvidia deal reshaped Groq

The backdrop matters. Nvidia’s $20 billion arrangement—widely described as a “not-acqui-hire”—licensed Groq’s custom inference silicon and absorbed its top leadership without a formal acquisition. That left Groq with its technology rights intact but stripped of the executives who built it. The company has since re-staffed and leaned into software and cloud services rather than competing head-on as a chip vendor against the same partner that just licensed its designs.

You can read TechCrunch’s reporting on how Groq confirmed the $650M raise and re-staffed after the Nvidia deal for the full breakdown.

Why investors still want in

The bet behind this Groq funding is that demand for fast, low-cost AI inference will keep outrunning supply. As more applications move from training breakthroughs to everyday production use, the companies that can serve tokens cheaply and quickly stand to win. Groq’s specialized architecture was always pitched as faster and more efficient than general-purpose GPUs for inference, and a neocloud model lets it monetize that without shipping hardware to every customer.

The takeaway

Surviving a talent raid by the most powerful company in AI and then raising $650 million is no small feat. The new Groq funding gives the startup a real runway to prove that an independent inference cloud can coexist with—and even feed off—the Nvidia ecosystem. Whether it can hit its 200-megawatt target on schedule will tell us a lot about how crowded the inference market is about to get.