Nvidia Took Groq's Founders — Groq Just Raised $650M

Groq confirmed a $650 million funding round on June 22, and the context matters as much as the number. In December 2025, Nvidia paid roughly $20 billion for a non-exclusive license to Groq's inference technology — and in the same move recruited Groq's founder Jonathan Ross (formerly of Google) and president Sunny Madra. Nvidia then used that license to launch its own inference product, the Nvidia Groq 3 LPX, at its March 2026 GTC conference. Six months later, Groq has raised $650 million, hired a new executive team, and is growing.

The new leadership includes Alan Rice as COO (previously at xAI and Meta), Sinclair Schuller as CTO (co-founder of Nuvalence), and Rakesh Malhotra as CPO (a Microsoft cloud veteran). The company is leaning into what it calls its neocloud business: a cloud inference service running 13 data centers across North America, Europe, the Middle East and Asia-Pacific. The numbers: more than 5 million developers, thousands of AI companies, and trillions of tokens processed each week.

What makes Groq different from other AI cloud providers

Groq's Language Processing Units (LPUs) were designed from scratch for one purpose: running a trained AI model as fast as the hardware allows. Standard cloud AI runs inference on GPUs originally built for training — powerful, but not purpose-built for the serve path. LPUs flip that. The result is that Groq's API has historically returned responses measurably faster than GPU-based alternatives. For most casual use cases the difference is minor. For latency-sensitive tasks — voice assistants, real-time analysis, AI agents that need to call a model mid-task without making a user wait — the gap is real and noticeable.

What the raise signals

The clearest takeaway is that Groq's cloud inference business is working on its own. The company lost its founder and a key executive to the biggest chip company in the world, watched Nvidia build a competing product using its own technology, and still grew its developer base and weekly token volume to where investors put in $650 million. That's not a startup propped up by hype — it's a provider with real usage.

The new leadership team's background — Meta, xAI, Microsoft cloud — points toward a push for broader enterprise access rather than the original hardware sales model. For developers and builders, that means Groq's API is likely to stay available and expand, not wind down.