Neocloud

Thunder Compute

GPU-over-network virtualization that undercuts nearly everyone on A100 instances.

Virtualizes GPUs over the network so instances share physical cards efficiently; the result is A100 pricing well below typical neocloud rates, with the trade-off of virtualization overhead on some workloads.

Visit Thunder ↗

GPU pricing on Thunder

On-demand list per GPU-hour, last reviewed 2026-09. Methodology.

GPU $/GPU-hr vs cheapest Notes
A100 40GB $0.66 cheapest Virtualized instances
A100 80GB $1.09 1.7× Virtualized instances

Strengths

  • Aggressively low A100 pricing
  • Per-minute billing, fast start

Watch for

  • Virtualized access, benchmark your workload
  • Young platform

Alternatives

CoreWeave

The flagship neocloud: Kubernetes-native GPU fleets built for training clusters, now public company.

Lambda

The ML researcher's default: one-click clusters, sane pricing, no quota theater.

RunPod

Community and secure cloud tiers, per-second billing, and serverless GPU endpoints.