Tracking 12 providers · 135 observations

Live Compute Pricing
GPU, AI & Data Centers

The real-time pricing layer for AI compute. Track live cloud pricing for H100, Blackwell, and next-generation GPUs across major providers. Built for operators, data center buyers, and investors who need visibility into compute markets.

Last updated: 2026-03-21T02:14:38.244Z

Live Compute Price Aggregation

Compute pricing fetched every 60 seconds from provider APIs across GPUs, data centers, and AI infrastructure.

Full Source Transparency

Every price is traceable to an exact API endpoint, parser, and timestamp. No black-box estimates.

Historical Price Tracking

Track how AI compute pricing moves over time — from H100 to Blackwell and future GPU generations.

Cross-Provider Benchmarking

Compare normalized compute costs across cloud providers, form factors, and spot vs on-demand markets.

Real-Time GPU Pricing

Live compute pricing indices across GPU generations — H100, Blackwell, and future architectures. Updated every 60 seconds.

Hopper (H100)

$2.3455 sources
View details
H100 Compute Pricing — Live Index
$2.34per GPU-hour
Loading chart data...

Blackwell (B200 / GB200)

$4.9011 sources
View details
Blackwell Compute Pricing — Live Index
$4.90per GPU-hour
Loading chart data...

How We Build the Compute Pricing Index

Indices are computed only from live API sources — providers whose compute prices are fetched dynamically every 60 seconds. Static published rate cards are displayed for reference but excluded from index computation.

Every price observation is weighted by confidence score and recency. The top and bottom 10% by price are trimmed. Per-provider weight is capped at 25% to prevent single-source dominance.

Every price is traceable to an exact source — API endpoint, parser version, and observation timestamp.

View Full Methodology

Explore Compute Pricing Data

Dive into detailed provider-level compute pricing, AI infrastructure cost analysis, and GPU spot market trends.