Analysis
Groq has raised a new round led by investment firm Disruptive, with Nvidia participating, according to TechCrunch. The numbers behind the reset:
- New round -- $350 million raised
- New valuation -- $3.5 billion
- Prior valuation (September 2025) -- $6.9 billion
- Nvidia licensing deal that triggered the reset -- $20 billion, to license Groq's technology and hire founder and CEO Jonathan Ross along with much of the company's top technical talent, an arrangement widely described at the time as a "not-acqui-hire"
A Groq spokesperson told TechCrunch the company does not consider this a down round, framing it instead as a fresh valuation for the "post-Nvidia-licensing-deal version of Groq" -- a company that has re-staffed and repositioned itself as a neocloud, selling GPU-based inference capacity rather than competing head-on with Nvidia on chip design. Groq now operates 13 data centers across North America, Europe, the Middle East and Asia Pacific, serving more than 6 million developers. Pulse previously covered Groq's LPU-based challenge to Nvidia before the licensing deal reshaped the company's strategy entirely.
“Groq now operates 13 data centers across North America, Europe, the Middle East and Asia Pacific, serving more than 6 million developers.”
From chip challenger to Nvidia's tenant
The pivot is a notable strategic retreat. Groq spent years positioning its LPU architecture as a genuine alternative to Nvidia GPUs for low-latency inference, and the Nvidia licensing deal effectively ended that chapter -- Groq's own technology now generates revenue for Nvidia rather than competing against it. What is left is a company running inference infrastructure at scale, competing against CoreWeave, Lambda, Together AI and Fireworks in a neocloud market that Nvidia itself is now financing on both sides of the table.
This is Groq's second raise in under three months, following a $650 million round in June, bringing total fresh capital to $1 billion during a period the company also lost its founding technical team to Nvidia. The pattern -- rapid re-staffing, rapid re-capitalization, valuation reset -- is becoming a template for AI chip startups that can't outrun Nvidia's roadmap: license the technology, take the check, and pivot to selling compute instead of silicon.
