Skip to content
Varni
Request capacity
Varni
Request GPU Capacity

01: Compute

Compute stack

  1. Workloads

    • AI training
    • AI inference
    • Fine-tuning
    • HPC
    • Production AI
    View workloads

  2. Managed compute

    Later phase

    • Serverless Inference
    • Kubernetes as a Service
    • Notebook as a Service
    • Slurm as a Service
    • Object Storage

  3. Bare metal

    Phase 1

    • Bare Metal GPU Compute
    • Training Clusters

  4. GPU capacity

    • Scheduled

      NVIDIA B300

      12 nodes · October 2026

    • Scheduled

      NVIDIA B300

      130 additional nodes · December 2026

    • Sourcing

      Additional GPU Capacity

      Sourcing against customer requirements

    View GPU capacity

02: Commercial

Commercial terms

Pay-as-you-goPay for GPU capacity as you use it. Terms confirmed per requirement.
Reserved capacityCapacity reserved for an agreed term. Terms confirmed per requirement.
Long-term commitmentCapacity over an agreed longer-term arrangement. Terms confirmed per requirement.
Enterprise / customCustom agreements for enterprise requirements. Terms confirmed per requirement.

Pricing: provided on request.