Karman’s AI-powered power management software has been implemented at a NextGen Cloud data center in Montreal, Canada on July 2, 2026. The firm claims that the application allowed 16 Nvidia B300 servers to function within the energy budget designated for ten servers. The company stated that this development provided 60% additional compute capacity and token throughput, while retaining inference quality and reliability.
Karman stated that the technology could maintain its performance across a variety of workloads, such as those with consistent high-power demands. The first collaboration was revealed in March, with the goal to enhance the AI compute resources within NexGen’s existing data centers, optimizing usage in a grid-limited setting.
Following the successful presentation, the companies have expanded the software deployment throughout NexGen’s European cloud facilities. The partners claim that, as the need for AI computing keeps growing, efficiency is becoming equally crucial as capacity. Karman’s success has proven that we can provide considerably increased computational power from our current infrastructure without needing extra allocated resources, according to Chris Starkey, CEO of NexGen Cloud.
This presents an enticing chance to enhance infrastructure efficiency, expedite implementation, and ultimately provide superior value to our clients. We are enthusiastic about further developing these outcomes as we broaden our collaboration throughout Europe. Karman’s platform employs AI technology to offer real-time visibility to grid operators and data center companies.
