Test report DSG-3972 · Rev E · tested October 30, 2026
AI Datacenter InfrastructureDevice under test
Karman software runs 16 Nvidia B300 servers on ten-server power budget
Karman's AI power orchestration software ran 16 Nvidia B300 servers on a ten-server power budget at a NextGen Cloud site in Montreal, adding 60% more compute capacity.
- Read
- 3 min
- Words
- 510
- Node
- 10nm
- Operator
- Amara Osei
Spec summary
- Karman's software ran 16 Nvidia B300 servers within a power budget provisioned for ten at NextGen Cloud's Montreal data center.
- The pilot delivered 60 percent more nameplate compute capacity and token throughput while maintaining inference quality.
- Karman raised $60.3 million in Series C funding in June, led by Renown Capital Partners with participation from Quanta, Nvidia and Keyframe Capital.
- The company rebranded from Utilidata this month and partnered with Legrand on data center power distribution integration.
- The rollout now extends across NextGen Cloud's European cloud operations.
Sixteen Nvidia B300 servers ran on a power budget provisioned for ten at NextGen Cloud's Montreal data center, according to Karman, whose AI-enabled power orchestration software delivered 60 percent more nameplate compute capacity and token throughput in the pilot.
The company reported that the deployment maintained inference quality and reliability across a range of workloads, including sustained high-power jobs. Karman announced the pilot's completion this week.
What does the pilot change for grid-constrained data centers?
Karman and NextGen Cloud announced their initial partnership in March 2025 with one stated goal: increase available AI compute capacity within NextGen's already-deployed data center footprint, maximizing capacity in a grid-constrained environment.
Following the demonstration, the two companies have extended the software rollout across NextGen's European cloud operations.
Chris Starkey, CEO of NexGen Cloud, said: "As demand for AI compute continues to accelerate, efficiency is becoming just as important as capacity. The results we've achieved with Karman demonstrate that we can deliver significantly more compute from our existing infrastructure without requiring additional provisioned power."
Starkey added: "This creates a compelling opportunity to improve infrastructure utilization, accelerate deployment, and ultimately deliver greater value to our customers. We're excited to build on these results as we expand our partnership across Europe."
The takeaway: the pilot turns a power constraint into additional compute without new grid connections.
How does the platform work?
Karman's platform uses AI to give grid operators and data center firms real-time visibility. It identifies and manages distributed energy resources, which the company claims enhances resiliency and offsets future infrastructure investments. The platform integrates with smart meters and provides AI solutions for utility companies and their customers.
Karman built the platform on a custom version of the Nvidia Jetson Orin Nano, an edge AI platform that the company says provides 100 times the processing power of other meter-based solutions. The platform can connect to millions of endpoints and coordinate them to monitor and react to real-time conditions.
Josh Brumberger, Karman CEO, said: "The fastest, low-cost way to scale AI today is by getting more compute out of the infrastructure already built."
Brumberger added: "Data centers hold power in reserve for worst-case scenarios that rarely happen. Karman turns that stranded capacity into usable compute. We're excited to bring this value to NexGen Cloud and their customers."
The mechanism in plain terms: reserve power held for rare peaks gets converted into running compute.
What else has Karman done this year?
The company rebranded earlier this month; it previously operated as Utilidata. Also this month, Karman partnered with French industrial group Legrand to integrate its software into Legrand's data center power distribution product lines.
In June, the firm raised $60.3 million in Series C funding led by Renown Capital Partners, with participation from Quanta, Nvidia, and existing investor Keyframe Capital.
Karman now carries a rebranded name, a European rollout, a Legrand integration and fresh capital into a market where power, not land, sets the ceiling on AI compute.
via Data Center Dynamics (Source)
More from Amara Osei
Show full bio
Staff writer covering money, funding and market shifts around hardware layer of artificial intelligence at Die Signal.
507 articles
Same lot · LOT-C204
- DSG-156045nmNvidia, Google and Emerald AI launch flexible data center consortium
- DSG-823214nmUpscale AI launches platform to interconnect rival-vendor AI chips
- DSG-112814nmOracle Lines Up Nvidia, Google and Cerebras AI Chips for Its Cloud
- DSG-16975nmNvidia's green-power push targets grid bottlenecks, revenue shield
- DSG-199314nmSWI Compute launches with 3.6GW power bank, $1bn Nvidia order