Groq secures $350M for Nvidia-based inference cloud expansion
Groq has raised $350 million to expand its neocloud business, which provides Nvidia GPU systems and AI infrastructure services for training and inference. Investment firm Disruptive led the financing, with planned participation from Nvidia, and the round values Groq at $3.5 billion.
The company said the capital will support customers seeking medium and larger clusters of Nvidia accelerated computing. Groq now operates 13 data centres across North America, Europe, the Middle East and Asia Pacific, serving more than 6 million developers, enterprises and AI-native companies.
A new phase after the Nvidia licensing deal
The $3.5 billion valuation is lower than Groq's $6.9 billion valuation last September. A company spokesperson said it should not be treated as a down round, describing it instead as a new valuation for the post-Nvidia-licensing-deal version of the company.
Groq had focused on its own LPUs, or language processing units, designed to compete with Nvidia in inference: the compute used to run AI workloads in real time. After Nvidia hired founder and chief executive Jonathan Ross and other senior talent through a licensing deal, Groq shifted from a pure chipmaker towards operating cloud and data-centre infrastructure based on Nvidia systems.
The change follows Groq's $650 million inference cloud financing, which marked the company's $650 million financing in June to begin its move into inference cloud operations. Groq is now an Nvidia customer while building capacity for businesses that need infrastructure for both model training and inference.
Capacity growth and neocloud economics
Groq intends to increase its available power capacity from 54 megawatts to more than 200 megawatts by 2027. Alex Davis, Groq chairman and CEO of Disruptive, said the company is building an AI inference cloud and called inference the largest and most critical layer of AI infrastructure.
The wider neocloud market is also being shaped by Nvidia's role as GPU supplier and investor. Nvidia systems power clouds run by CoreWeave, Lambda and Nebius, while Nvidia has invested billions in some capacity builders. CoreWeave has reported strong second-quarter revenue growth and won contracts with Meta and Anthropic, but investors have also focused on high capital expenditure, debt, hardware depreciation and the path to free cash flow.
Groq's financial results remain private, so the long-term profitability of its expanded model is not yet visible. For businesses procuring AI infrastructure, the practical implication is to evaluate cluster availability and workload fit alongside pricing and the economics of providers making large, hardware-intensive capacity commitments.