Cloud GPU Pricing — Cost Intelligence for H100, A100 & B200
GPU pricing stopped being a stable lookup table. We track 15 providers, every major GPU, and every pricing tier independently, so your team has a neutral view before committing budget.
$2.50
Cheapest H100 /GPU/hr
Hyperstack · 64% below AWS
$1.56
Cheapest H100 spot /GPU/hr
Hyperstack H100 NVLink 80GB
15
Providers Tracked
Marketplace rates shown separately
48
GPU Models Tracked
Distinct models in this view
Tools
GPU price arbitrage by region
Pick a GPU model and see every tracked provider's on-demand rate, grouped by region and sorted cheapest to most expensive.
Cheapest
Azure in US East
$0.90/GPU-hr
Most expensive tracked
Lambda in US West
$1.29/GPU-hr
1.4x price difference ($0.39/GPU-hr gap)
US East
| Provider | Instance | $ / GPU-hr |
|---|---|---|
| Azure | NV18ads_A10_v5 | $0.90 |
NV18ads_A10_v5
US West
| Provider | Instance | $ / GPU-hr |
|---|---|---|
| Lambda | gpu_1x_a10 | $1.29 |
gpu_1x_a10
Prices as tracked; confirm current rates on the provider's page before procurement.
| Trend | Actions | Deploy | ||||
|---|---|---|---|---|---|---|
LiumCheapest RTX A4000 · 16 GB | $0.12 — vs AWS | —Live range: $0.12–$0.12/GPU-hr | — | Deploy → | ||
Hyperstack A4000 16GB · 16 GB | $0.15 — vs AWS | $0.1220% off | $0.11 | Deploy → | ||
Lium RTX 3090 · 24 GB | $0.18 — vs AWS | —Live range: $0.18–$0.60/GPU-hr | — | Deploy → | ||
Lium RTX 5060 Ti · 16 GB | $0.20 — vs AWS | —Live range: $0.20–$0.20/GPU-hr | — | Deploy → | ||
RunPod RTX A5000 24GB · 24 GB | $0.27 — vs AWS | $0.1641% off Community Cloud (shared): $0.16/GPU-hr | — | Deploy → | ||
Thunder Compute RTX A6000 48GB · 48 GB | $0.35 — vs AWS | — | — | Deploy → | ||
Lium RTX 4090 · 24 GB | $0.40 — vs AWS | —Live range: $0.40–$0.58/GPU-hr | — | Deploy → | ||
Lium RTX A6000 · 48 GB | $0.42 — vs AWS | —Live range: $0.42–$0.42/GPU-hr | — | Deploy → | ||
RunPod A40 48GB · 48 GB | $0.49 — vs AWS | $0.3529% off Community Cloud (shared): $0.35/GPU-hr | — | Deploy → | ||
RunPod L4 24GB · 24 GB | $0.49 — vs AWS | $0.4410% off Community Cloud (shared): $0.44/GPU-hr | — | Deploy → | ||
RunPod RTX 3090 24GB · 24 GB | $0.50 — vs AWS | $0.2256% off Community Cloud (shared): $0.22/GPU-hr | — | Deploy → | ||
Hyperstack A6000 48GB · 48 GB | $0.50 — vs AWS | $0.4020% off | $0.35 | Deploy → | ||
Lium RTX 5090 · 32 GB | $0.50 — vs AWS | —Live range: $0.50–$0.75/GPU-hr | — | Deploy → | ||
RunPod RTX A6000 48GB · 48 GB | $0.53 — vs AWS | $0.3338% off Community Cloud (shared): $0.33/GPU-hr | — | Deploy → | ||
RunPod Pro 6000 MIG 24GB · 24 GB | $0.59 — vs AWS | $0.5015% off Community Cloud (shared): $0.50/GPU-hr | — | Deploy → | ||
Lambda Quadro RTX 6000 24GB · 24 GB | $0.69 — vs AWS | — | — | Deploy → | ||
Lium RTX 6000 Ada Generation · 48 GB | $0.70 — vs AWS | —Live range: $0.70–$1.04/GPU-hr | — | Deploy → | ||
RunPod RTX 4090 24GB · 24 GB | $0.74 — vs AWS | $0.3454% off Community Cloud (shared): $0.34/GPU-hr | — | Deploy → | ||
Enverge DGX Spark · 128 GB | $0.75 — vs AWS | — | — | — | ||
Enverge DGX Spark · 128 GB | $0.75 — vs AWS | — | — | — | ||
Lambda Tesla V100 16GB · 128 GB (8×16) | $0.79 — vs AWS | — | — | Deploy → | ||
Thunder Compute L40 48GB · 48 GB | $0.79 — vs AWS | — | — | Deploy → | ||
RunPod L40 48GB · 48 GB | $0.82 — vs AWS | $0.6916% off Community Cloud (shared): $0.69/GPU-hr | — | Deploy → | ||
Lium L40 · 48 GB | $0.82 — vs AWS | —Live range: $0.82–$0.82/GPU-hr | — | Deploy → | ||
Enverge DGX Spark · 256 GB | $0.82 — vs AWS | —CALC: $1.65 / 2 GPUs | — | — |
Reserved-only rates
These capacity-reservation rates have no pay-as-you-go tier, so they are listed here and not compared with on-demand prices above. Rates are provider-reported, per GPU, last tracked as shown.
| Provider | GPU | Region | Reserved $/GPU-hr | Term | Last tracked |
|---|---|---|---|---|---|
| GCP | B200 192GB | us-central1 | $16.11 | Reservation required (Flex-start) | 2026-09-11 |
| FluidCore | B300 288GB | US | $5.20 | Reserved (term not specified) | 2026-10-05 |
| FluidCore | B300 288GB | Europe | $5.20 | Reserved (term not specified) | 2026-10-05 |
| FluidCore | B300 288GB | India | $5.20 | Reserved (term not specified) | 2026-10-05 |
| FluidCore | H200 141GB | India | $3.50 | Reserved (term not specified) | 2026-10-05 |
Use Spot for Training
Save 60–70% on training runs with checkpointing. GCP Spot offers up to 70% discount on A100/H100 instances.
Best for: Fault-tolerant training with checkpoints
Reserved for Inference
1-year commitments save 35–40% for always-on inference endpoints. Azure and AWS offer the deepest reserved discounts.
Best for: Production inference workloads
Lambda/Nebius for Value
GPU-specialist clouds offer 2–4× lower per-GPU pricing than hyperscalers. Nebius leads on committed pricing; Lambda leads on on-demand.
Best for: Pure GPU compute without cloud services
Provider Directory
Hyperscaler. Capacity Blocks and Savings Plans for committed GPU capacity.
Hyperscaler. Offers NVIDIA GPUs and Google TPUs.
Hyperscaler. Offers NVIDIA and AMD Instinct GPU instances.
GPU-focused cloud with on-demand and reserved clusters.
GPU-focused cloud. Committed pricing is quoted per contract.
Secure Cloud (dedicated) and Community Cloud (shared) are separate tiers here.
GPU cloud with EU and US data centers.
Inference API and dedicated GPU clusters on one platform.
GPU cloud. Provider states no egress fees.
GPU cloud with European regions. Provider states no egress fees.
GPU cloud with per-minute billing.
GPU sandboxes for AI-generated code execution, billed per second.
Small-form-factor and workstation GPU hosting (UK, Portugal).
Marketplace. Rates are set by individual hosts and shown separately from dedicated-tier rates.
GPU node rental. Rates shown per node, normalized per GPU.
Prices vary by region and availability. Verify with the provider before procurement. Rows show when they were last tracked.