Groq Secures $350M to Fuel Pivot From AI Chips to Neocloud
AI infrastructure startup Groq has secured $350 million in new funding led by investment firm Disruptive, with planned participation from Nvidia. The new capital values the company at $3.5 billion—a steep drop from its $6.9 billion valuation last September, but a key marker for its ongoing shift from a custom chipmaker into a neocloud provider.
The valuation reset comes on the heels of a massive $20 billion licensing deal in late 2025, which saw Nvidia acquire rights to Groq’s inference tech while hiring away founder Jonathan Ross and key engineering leadership. Rather than continuing to compete directly with Nvidia’s dominance via its proprietary Language Processing Units (LPUs), Groq is aggressively rebuilding itself as a specialized cloud operator. The company now rents out compute clusters—powered primarily by Nvidia GPUs alongside its legacy hardware—to deliver fast AI model inference at scale.
The new funding will expand Groq’s physical footprint. Building on a $650 million raise from June, Groq currently operates 13 data centers worldwide with 54 megawatts of capacity. It aims to scale past 200 megawatts by 2027 to serve its growing user base of over 6 million developers and enterprise clients.
Groq’s reinvention underscores a broader shift across the AI sector. While competing head-to-head with Nvidia on hardware design remains a grueling uphill battle, supplying raw, managed compute capacity to inference-hungry developers offers a far more sustainable path to near-term revenue. Expect Groq to continue leaning hard into infrastructure expansion as it carves out its space alongside rival neoclouds like CoreWeave.