Expanse
United Kingdom · AI
Predictive infrastructure software that forecasts GPU, memory, runtime and failure risk before AI or HPC jobs run, helping teams use existing clusters more efficiently.
Plan compute before jobs enter the queue
Expanse connects to an existing SLURM, Kubernetes or Nomad cluster and learns from its workload history. Before a job runs, it predicts GPU, memory and runtime needs and presents the recommendation with confidence and supporting evidence, so teams can avoid both under-sized jobs that fail and oversized requests that leave capacity idle.
Keep the feedback loop inside the cluster
After execution, Expanse can diagnose failures from code, logs and metrics; observed outcomes improve later predictions. The models and management layer run within customer-controlled infrastructure, keeping code and telemetry on the cluster while fitting into existing submission and scheduling workflows.
- Product
- Predictive compute infrastructure
- Customer
- Teams operating AI and HPC clusters
Funding history
1 round · 1 with disclosed amounts| Announced | Round | Amount | Investors |
|---|---|---|---|
| Seed | €4.5M | Crane Venture Partners — LeadAIP Seed — ParticipantPXN Ventures — Participant |