GPU Rental Prices
SearchCtrl K

H100 rental prices, 2026-10

Across 23 matched firm-price configurations, prices changed by 3.5% between 2026-10-01 and 2026-10-13. 7 configurations changed price.

Checked 13 October 2026. This is a retrospective calculation from the recorded rates. This report covers 13 recorded days; it does not assume observations for missing days or the unfinished part of a month.

At the end of the period, H100's median firm price was $3.33/GPU-hour across 24 providers. The hyperscaler median was 117.8% above the other-cloud median.

Largest fall · Scaleway H100 SXM
1.5% lower

$3.63 → $3.57/GPU-hour · 2026-10-01 to 2026-10-13

◀ Lower pricesHigher prices ▶
Scaleway H100 SXM▼ 1.5%
Scaleway H100▼ 1.5%
Spheron H100▲ 3.8%
DataCrunch H100 SXM▲ 12.7%
RunPod H100 SXM▲ 14.3%
Jarvislabs H100 SXM▲ 29.7%
GMI Cloud H100▲ 30.0%
Like-for-like firm-price changes · 2026-10-01 to 2026-10-13. Bar length shows the percentage change from the zero line.

Prices that changed

Provider / GPU variantFirst $/GPU-hourLast $/GPU-hourChange
Scaleway · H100 SXM$3.63$3.57-1.5%
Scaleway · H100$3.25$3.21-1.5%
Spheron · H100$2.65$2.753.8%
DataCrunch · H100 SXM$3.63$4.0912.7%
RunPod · H100 SXM$3.49$3.9914.3%
Jarvislabs · H100 SXM$2.69$3.4929.7%
GMI Cloud · H100$2.00$2.6030.0%

What this change measures

The headline is the geometric mean of end/start price ratios for configurations present at both endpoints. Matching requires the same provider, GPU variant, tier, minimum GPU count, and published instance and region. A provider entering or leaving the sample is not counted as a price change. No cause for a change is inferred from its price alone.

Reserved and spot rows are excluded from this firm-rate movement. See current tier discounts separately. Compare current H100 offers.

Whole-market report · All monthly reports · Price index

Compare your offers

Saved prices keep their original dates. Confirm the configuration, terms and availability at the source.

Select up to three offers to compare their rates and terms.

Type a GPU, a provider, a model or a question.