Groq Raises $350M, Shifts to Neocloud Services

Startup Groq secures $350 million in funding as it transitions from an AI chipmaker to a neocloud company offering potent GPUs and AI infrastructure services. The recent investment, led by Disruptive, with Nvidia planned to participate, values the company at $3.5 billion. Despite a lower valuation from its previous $6.9 billion, Groq sees it as establishing a new post-Nvidia-licensing-deal valuation. Previously focused on developing its own chips to compete with Nvidia, Groq pivoted to cloud and data center services after losing key team members.

The company's intention is to expand from 54 megawatts to over 200 megawatts by 2027, currently operating 13 data centers across various regions and serving millions of developers, enterprises, and AI-native companies. The fresh funds will support the usage of medium and larger sized clusters of Nvidia accelerated computing for training and inference. Groq aims to become a leading AI inference cloud, with the belief that inference will be the most crucial layer of AI infrastructure.

However, concerns remain about the profitability of neocloud businesses in the long term. Groq's financials are currently private, but the pivot places the company within Nvidia's AI infrastructure ecosystem. Nvidia's involvement in supplying GPUs to neoclouds like CoreWeave, Lambda, and Nebius, as well as investing in their capacity expansion, indicates a common trend in the industry. TechCrunch has sought further details from Groq.