Groq, an AI chip startup focused on inference hardware, has secured $750 million in new funding, bringing its valuation up to $6.9 billion. The fresh capital comes amid growing investor interest in chips designed not just for training massive models, but for running them efficiently in production.
The funding round was led by the firm Disruptive, with major contributions from investors such as BlackRock, Neuberger Berman, and Deutsche Telekom Capital Partners. Existing backers including Samsung, Cisco, and Altimeter also participated. With this new injection of capital, Groq’s valuation has more than doubled compared to a year ago, when it was valued around $2.8 billion.
Groq differentiates itself by developing what it calls LPUs (Language Processing Units)—hardware optimized for inference workloads. These chips aim to deliver fast responses for AI models with lower latency and cost, especially in settings where real-time performance matters. Groq offers its hardware both through cloud services and on-premises clusters, making it accessible for a variety of users including enterprises trying to reduce dependence on generic GPU-based inference.
In recent years, demand for inference hardware has surged. The industry has been shifting focus from raw training capacity to models that can serve predictions reliably and at scale. As more AI applications move into user-facing areas—chat, recommendation systems, real-time decision making—the need for efficient inference becomes more pressing. Groq’s work sits squarely in that trend.
The fresh funds are expected to accelerate Groq’s product development, expand its manufacturing and deployment capacity, and enhance support for customers using its hardware. Industry watchers see this raise as yet another sign that AI infrastructure is becoming more competitive, with challengers to established GPU giants like Nvidia growing stronger.
