Groq Raises $650M to Build the World's AI Inference Cloud After Selling LPU IP to Nvidia
Summary: Six months after licensing its inference technology to Nvidia in a $20B deal, Groq announced a $650M funding round to scale its neocloud business to 200 MW by 2027, with 13 data centers already operating across four continents.
---
AI inference startup Groq confirmed a $650 million funding round on Monday, roughly six months after it entered into a non-exclusive licensing agreement with Nvidia that saw its founder and key talent move to the GPU giant.
The round was led by Disruptive and Infinitum, with participation from existing investors. Groq didn't disclose its new valuation — it was last valued at $6.9 billion following its September 2025 raise.
The Backstory: The Nvidia Deal That Reshaped Groq
In December 2025, Groq signed a non-exclusive licensing agreement with Nvidia. As part of the deal, founder and CEO Jonathan Ross (who helped create Google's TPU), president Sunny Madra, and several other employees joined Nvidia. Nvidia gained access to Groq's LPU (Language Processing Unit) technology and announced its own hardware cluster — the Nvidia Groq 3 LPX — at GTC in March 2026.
What happened next is unusual: Groq didn't wind down. It pivoted.
The Pivot: Neocloud Over Chips
With its LPU IP now licensed to Nvidia, Groq is doubling down on its inference cloud business. The company has grown its neocloud footprint to 13 data centers across North America, Europe, the Middle East, and APAC. It claims to serve over 5 million developers and thousands of AI companies, processing trillions of tokens each week.
The new capital will accelerate the fit-out of existing data centers with Groq's latest inference technology, including the new LPX system from Nvidia. Groq expects to scale toward 200 MW of capacity by the end of 2027.
New Leadership Team
Groq has been rebuilding its executive ranks:
• Alan Rice joins as COO, previously at xAI and Meta after a career in the U.S. Navy
• Sinclair Schuller comes on as CTO, founder of Apprenda and co-founder of Nuvalence
• Rakesh Malhotra joins as CPO, former Microsoft cloud veteran and Schuller's co-founder at Nuvalence
What This Means for AI Inference Pricing
Groq's continued independence is good news for developers who value inference speed. Groq's LPU-based inference has been competitive on latency, particularly for open-weight models like Llama and Mistral. With Groq now operating as a pure neocloud player rather than a chip designer, its incentives align more closely with AWS or Azure — maximizing throughput on its infrastructure. The 200 MW target suggests serious capacity expansion, which could drive inference prices down for GroqCloud users.
The Nvidia LPX system incorporating Groq's technology also means that Groq's architecture is effectively becoming a platform both companies sell. Nvidia handles the enterprise direct sales; Groq handles the cloud-native developer market.
The Takeaway
Groq's path is unusual in the AI chip space. Most startups that license their core IP to Nvidia don't survive as independent companies. Groq is betting that its neocloud — not its silicon — is the lasting business. With 5 million developers already on the platform and $650M to deploy, they have a real shot at proving it.
Sources:
• Groq — Official Announcement