Expanse, a compute optimization platform predicting hardware requirements for AI and high-performance computing workloads, has raised $5.3 million in funding led by Crane Venture Partners. PXN Ventures and angel investors, including former DeepMind researchers and AI infrastructure leaders, also participated.
The funding will support engineering growth and deployment across AI labs, quantitative finance, life sciences and research computing.
Predicting Compute Needs Before Workloads Run
Expanse addresses inefficient hardware allocation across large computing clusters, where teams often reserve more capacity than workloads require to avoid failures caused by insufficient resources. In one cluster, Expanse identified nearly $8 million of idle compute capacity in a single month.
Expanse analyzes workload code, historical cluster data and available hardware to predict resource requirements before GPUs or other computing resources are allocated. The platform supports Slurm, Kubernetes, major cloud providers, on-premise infrastructure and hybrid environments.
Unlocking Existing Computing Capacity
Growing demand for GPUs, long hardware delivery times and power constraints for data centres are increasing pressure on organizations to improve utilization of existing infrastructure. Expanse aims to release unused capacity by reducing unnecessary hardware reservations while lowering the risk of workloads failing because of insufficient resources.
The platform operates within customer environments, keeping code and data inside existing infrastructure. Unlike monitoring tools focused on historical usage, Expanse makes resource predictions before workloads begin.
Expanding Across Compute-Intensive Industries
The $5.3 million investment will primarily support engineering and wider adoption across compute-intensive sectors. Expanse plans to deploy the platform across additional clusters serving AI development, financial modeling, scientific research and life sciences.