QumulusAI Scales GPUs, but Powered Capacity Sets the Pace

An AI infrastructure provider more than tripled its deployed GPU fleet in the second quarter while signing new customer contracts, demonstrating how companies are building out AI compute capacity. The company's growth reveals that success in this market depends on securing available powered data center space that can be deployed on customer timelines, rather than waiting for large-scale data center projects to be completed. Newer GPU deployments are generating higher revenue per megawatt, but this requires substantial capital investment in both GPU purchases and data center infrastructure. The broader industry context shows multiple specialized providers are pursuing modular deployment approaches to bring AI compute online faster, with execution speed becoming a key competitive factor.