Skip to content
regularly tracked, last cycle August 2026 · some rates updated outside cycle

Cloud GPU Pricing — Cost Intelligence for H100, A100 & B200

GPU pricing stopped being a stable lookup table. Supply constraints and new hardware releases mean prices move meaningfully quarter to quarter. We track them independently — across 11 providers, every major GPU, and all pricing tiers — so your team has a neutral view before committing budget.

✓ 11 cloud providers✓ H100, H200, A100, L40, MI300X✓ On-demand & reserved rates✓ Spot & interruptible pricing

Cloud GPU Pricing — Aug 2026

Compare Cloud GPU Costs

Independent cost intelligence across major GPU cloud providers. Prices are regularly tracked.

Prices regularly trackedSome "Rent Now" links are affiliate links — we earn a small commission at no cost to you. Price data and rankings are never influenced by commercial relationships.regularly tracked, last cycle August 2026 · some rates updated outside cycle

$2.19

Cheapest H100 /GPU/hr

Thunder Compute · 68% below AWS

$0.12

Best Spot Deal

Hyperstack A4000 16GB

10

Providers Tracked

Major cloud platforms

10+

GPU Types

B300, H200, H100, MI300X, TPU…

Tools

Reserved vs On-Demand break-even

See how many months until your reserved commitment pays off versus on-demand pricing.

Commitment Analyzer

Reserved vs on-demand break-even

Model the full-term cost of a GPU commitment versus on-demand at your expected utilization, across three market-movement scenarios.

Total cost over 12 months

Market −20%On-demand wins

On-demand total

$90,753.60

Reserved total

$105,120.00

On-demand saves $14,366.40

FlatReserved wins

On-demand total

$113,442.00

Reserved total

$105,120.00

Reserved saves $8,322.00

Market +20%Reserved wins

On-demand total

$136,130.40

Reserved total

$105,120.00

Reserved saves $31,010.40

Cost vs utilization (flat scenario)

Monthly cost at each utilization level

730 hr/mo max
On-demand
Reserved
Break-even

Verdict

At 70% utilization, a 12-month commitment saves $8,322.00 — but if market prices fall 20%, on-demand wins by $14,366.40.

Estimates only. Reserved rates assume locked-in pricing for full commitment term. Confirm current rates on the provider's pricing page before procurement.

per nodeper GPU
Provider:
GPU:
55 instances · prices shown per hour · normalized per GPU
Cloud GPU instance pricing across AWS, GCP, Azure, Lambda, CoreWeave, RunPod, Nebius, Together AI, Crusoe, and Hyperstack
Provider GPU VRAM On-Demand $/GPU-hr Spot $/GPU-hr 1Y Reserved $/GPU-hr vs AWSHistAlertWatchDeploy
HyperstackA4000 16GB16 GB
$0.15
$0.1220% off
$0.11Deploy →
Thunder ComputeRTX A6000 48GB48 GB
$0.35
Deploy →
HyperstackA6000 48GB48 GB
$0.50
$0.4020% off
$0.35Deploy →
GCPTPU v5e128 GB HBM
$0.60
$0.38Deploy →
Thunder ComputeL40 48GB48 GB
$0.79
Deploy →
Thunder ComputeL40S 48GB48 GB
$0.99
−51%Deploy →
HyperstackL40 48GB48 GB
$1.00
$0.8020% off
$0.70Deploy →
Thunder ComputeA100 80GB80 GB
$1.09
Deploy →
HyperstackA100 PCIe 80GB80 GB
$1.35
$1.0820% off
$0.95Deploy →
RunPodA100 80GB640 GB (8×80)
$1.49
$0.8145% off
Deploy →
GCPTPU v6e256 GB HBM
$1.56
$0.98Deploy →
GCPTPU v4128 GB HBM
$1.60
$0.96Deploy →
HyperstackA100 SXM 80GB80 GB
$1.60
$1.2820% off
$1.36Deploy →
HyperstackRTX Pro 6000 SE 96GB96 GB
$1.85
$1.4820% off
$1.30Deploy →
AWSL40S 48GB384 GB (8×48)
$2.00
$0.6966% off
$1.25baselineDeploy →
Thunder ComputeH100 PCIe 80GB80 GB
$2.19
−68%Deploy →
CoreWeaveL40S 48GB48 GB
$2.25
+12%Deploy →
CoreWeaveA100 80GB80 GB
$2.50
Deploy →
HyperstackH100 PCIe 80GB80 GB
$2.50
$2.0020% off
$1.75−64%Deploy →
HyperstackH100 NVLink 80GB80 GB
$2.60
$1.5640% off
$1.82−62%Deploy →
GCPTPU v5p384 GB HBM
$2.64
$1.66Deploy →
LambdaA100 80GB640 GB (8×80)
$2.79
Deploy →
RunPodH100 80GB640 GB (8×80)
$2.99
$1.9933% off
−57%Deploy →
HyperstackH100 SXM 80GB80 GB
$3.20
$1.9240% off
$2.72−53%Deploy →
CrusoeMI300X 192GB1536 GB (8×192)
$3.45
Deploy →
GCPTPU v7384 GB HBM3e
$3.56
$2.31Deploy →
AzureA100 80GB640 GB (8×80)
$3.56
$1.0770% off
$2.22Deploy →
NebiusH100 80GB640 GB (8×80)
$3.85
$2.15−44%Deploy →
CrusoeH100 80GB640 GB (8×80)
$3.90
−43%Deploy →
LambdaH100 80GB640 GB (8×80)
$3.99
$2.40−42%Deploy →
Together AIH100 80GB640 GB (8×80)
$3.99
$3.19−42%Deploy →
HyperstackH200 141GB141 GB
$3.99
$2.8030% off
$2.79−69%Deploy →
RunPodH200 141GB1128 GB (8×141)
$4.39
$2.7537% off
−66%Deploy →
NebiusH200 141GB1128 GB (8×141)
$4.50
$2.45−65%Deploy →
GCPA100 80GB640 GB (8×80)
$5.07
$1.5270% off
$3.17Deploy →
HyperstackB200 192GB192 GB
$6.00
$4.8020% off
$5.10−58%Deploy →
CoreWeaveH100 80GB80 GB
$6.16
−10%Deploy →
CoreWeaveH200 141GB141 GB
$6.31
−51%Deploy →
AWSH100 80GB640 GB (8×80)
$6.88
$2.5962% off
$4.81baselineDeploy →
AzureMI300X 192GB1536 GB (8×192)
$7.86
$2.3670% off
$4.94Deploy →
CoreWeaveB200 192GB192 GB
$8.60
−40%Deploy →
AzureMI250X 128GB512 GB (4×128)
$9.00
$2.7070% off
$5.50Deploy →
GCPH200 141GB1128 GB (8×141)
$10.85
$3.2071% off
$7.03−17%Deploy →
GCPH100 80GB640 GB (8×80)
$11.06
$3.3070% off
$6.90+61%Deploy →
AzureH100 80GB640 GB (8×80)
$11.06
$2.2180% off
$7.19+61%Deploy →
AzureMI300A 128GB512 GB (4×128)
$11.25
$3.3870% off
$7.00Deploy →
AzureMI325X 288GB2304 GB (8×288)
$12.25
$3.6770% off
$7.75Deploy →
AWSH200 141GB1128 GB (8×141)
$13.00
$4.7563% off
$8.50baselineDeploy →
AzureMI355X 288GB2304 GB (8×288)
$13.50
$4.0570% off
$8.75Deploy →
GCPB200 192GB1536 GB (8×192)
$13.75
$4.1370% off
$9.00−4%Deploy →
AzureB200 192GB1536 GB (8×192)
$14.00
$4.2070% off
$9.13−3%Deploy →
AWSB200 192GB1536 GB (8×192)
$14.38
$5.2563% off
$9.38baselineDeploy →
GCPB300 288GB2304 GB (8×288)
$17.75
$5.3370% off
$11.50−4%Deploy →
AzureB300 288GB2304 GB (8×288)
$18.13
$5.4470% off
$11.75−2%Deploy →
AWSB300 288GB2304 GB (8×288)
$18.50
$6.5065% off
$12.00baselineDeploy →

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 Intelligence

Cost · Reliability · Ecosystem rated 1–5
AWSMin: None
Cost
Reliability
Ecosystem

Most complete cloud. Premium pricing. Best SLA.

GCPMin: None
Cost
Reliability
Ecosystem

Best TPU access. JAX-native. Strong SLA.

AzureMin: None
Cost
Reliability
Ecosystem

Best AMD GPU selection. Enterprise-ready.

LambdaMin: None
Cost
Reliability
Ecosystem

Best $/GPU for pure compute. No ecosystem overhead.

CoreWeaveMin: 3 mo
Cost
Reliability
Ecosystem

Hyperscaler-grade infra at specialist pricing.

RunPodMin: None
Cost
Reliability
Ecosystem

Cheapest H100 on-demand. Community cloud model.

NebiusMin: None
Cost
Reliability
Ecosystem

Best committed pricing. EU data centers.

Together AIMin: None
Cost
Reliability
Ecosystem

Inference API + raw GPU in one platform.

CrusoeMin: None
Cost
Reliability
Ecosystem

Zero egress fees. OpenAI Stargate infrastructure partner.

HyperstackMin: None
Cost
Reliability
Ecosystem

Zero-egress European cloud with strong price-performance on H-series.

Thunder ComputeMin: None
Cost
Reliability
Ecosystem

YC-backed GPU cloud. Zero egress. Per-minute billing. Prices reflect GPU-over-TCP virtualized instances (1x–2x GPU configs). A100 80GB: $1.09/GPU-hr for a single GPU; multi-GPU configs are priced above that rate — 2× $2.18, 4× $5.96, 8× $11.92 (≈$1.49/GPU at 4×–8×). Do not assume linear scaling from the single-GPU rate. US + Canada (provider-stated, July 2026). regularly tracked · last tracked July 27, 2026.

Prices are approximate and vary by region and availability. Pricing reflects Aug 2026 estimates — always verify with provider pricing pages before procurement.