Market snapshotB200 $7.21 −2.6%
AI COMPUTE
TRACKER

01GPU cloud· as of 2026-07-06

What compute rents for

A venue-by-venue view of on-demand accelerator pricing. Pick a chip — the statistics, history, and quotes below all follow it.

H100 chip page →

H100 · market floor

$1.74/GPU·hr

Vast.ai · marketplace

H100 · median of tracked quotes

$3.95/GPU·hr

14 providers tracked

H100 · hyperscaler list

$6.88/GPU·hr

AWS (p5) on-demand

H100 SXM · on-demand price history

US$ per GPU-hour, quarterly midpoints by provider class

Price-history chart awaiting interactive loadStatic chart frame · exact values in table view

Indicative series compiled from provider list prices & archived snapshots (±15%). Current-quarter points reflect verified quotes.

H100 · current on-demand quotes

Sorted cheapest-first; use the column headers to re-sort.

14 venues quoting H100
Current H100 on-demand prices per GPU-hour by provider
Vast.ai — open source evidenceest — estimated / secondary-source figureMarketplace$1.74
Voltage Park — open source evidenceSpecialist$1.99
SF Compute — open source evidenceMarketplace$2.04
Hyperstack — open source evidenceSpecialist$3.20
RunPod — open source evidenceSpecialist$3.29
Nebius — open source evidenceSpecialist$3.85
Crusoe — open source evidenceSpecialist$3.90
Lambda — open source evidenceSpecialist$3.99
Together AI — open source evidenceSpecialist$3.99
CoreWeave — open source evidenceSpecialist$6.16
AWS (p5) — open source evidenceHyperscaler$6.88
Oracle OCI — open source evidenceest — estimated / secondary-source figureHyperscaler$10.00
Google Cloud (A3) — open source evidenceest — estimated / secondary-source figureHyperscaler$10.98
Azure (ND v5) — open source evidenceHyperscaler$12.29

Spot board · implied price vs venue quotes

Market-implied spot (Kalshi, JUL 17 2026 · 22:59 UTC) beside published on-demand list ranges by venue class (2026-07-06). One chip does not have one price — the dispersion is why indices exist.

Same GPU, 7.1× spread (H100)
Market-implied spot prices and venue-class quote ranges per chip
ChipImplied spot24hFront monthHyperscalerSpecialistMarketplace
B200HBM3e · 180GB$7.21−2.6%$6.67$14.24–$16.111$5.89–$8.602$4.693
H200HBM3e · 141GB$4.71+6.1%$4.41$7.91–$13.784$2.30–$6.315$2.896
H100SXM · 80GB$2.69+5.9%$2.62$6.88–$12.297$1.99–$6.168$1.74–$2.049
A100SXM4 · 80GB$1.05−5.4%$1.10$1.35–$2.7910
RTX 5090GDDR7 · 32GB$0.56−6.7%$0.56$0.9911$0.27–$0.3312
Provider quote sources12 notes
  1. 1 AWS (p6) $14.24 · Google Cloud (A4) $16.11
  2. 2 RunPod $5.89 · Hyperstack $6.00 · Modal $6.25 · Lambda $6.69 · Nebius $7.15 · Together AI $8.19 · CoreWeave $8.60
  3. 3 Vast.ai $4.69
  4. 4 AWS (p5en) $7.91 · Azure (ND v5) $13.78
  5. 5 FluidStack $2.30 · Hyperstack $3.99 · Crusoe $4.29 · RunPod $4.39 · Nebius $4.50 · CoreWeave $6.31
  6. 6 Vast.ai $2.89
  7. 7 AWS (p5) $6.88 · Oracle OCI $10.00 · Google Cloud (A3) $10.98 · Azure (ND v5) $12.29
  8. 8 Voltage Park $1.99 · Hyperstack $3.20 · RunPod $3.29 · Nebius $3.85 · Crusoe $3.90 · Lambda $3.99 · Together AI $3.99 · CoreWeave $6.16
  9. 9 Vast.ai $1.74 · SF Compute $2.04
  10. 10 Hyperstack $1.35 · RunPod $1.49 · Crusoe $2.30 · CoreWeave $2.70 · Lambda $2.79
  11. 11 RunPod $0.99
  12. 12 Salad $0.27 · Vast.ai $0.33