The dashboard tracks daily on-demand rental prices per GPU-hour across AWS, Azure, RunPod’s Secure and Community clouds, and Vast.ai’s marketplace, with figures dated November 11, 2026. Prices vary sharply by provider and model: an H100 SXM costs $6.88 on AWS and $12.29 on Azure, compared with $3.99 on RunPod Secure, $2.69 on Community, and a Vast.ai median of $2.94. For an A100 80GB, the listed prices range from Vast.ai’s $1.10 median to Azure’s $4.10. Newer high-end models are much costlier: B300 prices span $6.94 to $17.80 per hour, while B200 prices span $5.98 to $14.24. The page also shows daily price-history charts beginning November 2.
Prices are normalized to one GPU: when a cloud sells an eight-GPU machine, its hourly price is divided by eight. The comparison covers compute only; storage, networking, and support are excluded. AWS and Azure figures are regional on-demand Linux list prices, while RunPod distinguishes its own data centres from third-party community hosts. Vast.ai reports the median and lowest rentable offer, plus the number of offers, so marketplace prices can move more when listings are scarce. Large daily list-price changes are held until checked, and stale provider data retains its previous price. Reserved or committed-use discounts are not included, so the figures compare published on-demand rates rather than total costs for every buyer.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.