Hundredth RunCompute Buyer Desk
Menu
PricesLead timesChecklistHow it worksFAQRequest GPUs

H100 rental price per hour

The cheapest public on-demand H100 price we found is $1.68 per GPU-hour at Latitude.sh (80GB). The median is $3.99 and the priciest is $12.29, across 28 on-demand listings from 17 providers. Last checked .

Cheapest
$1.68
Latitude.sh
Median
$3.99
28 on-demand listings
Priciest
$12.29
On-demand, per GPU-hour
Providers
17
with an on-demand H100 price

Listings

Every H100 listing we track, 58 rows

US dollars per GPU-hour from public pages, cheapest on-demand first. Rows with no public price say why. Every row links to its source.

H100 prices by provider, with source link and date checked
Provider$ / GPU-hrVariantType and termRegionAvailabilitySourceChecked
Latitude.sh$1.6880GBOn-demandNot itemizedBare metalSource2026-11-01
Voltage Park$1.99HGXOn-demandNot statedPage: self-serve, up within 15 mins, no minimum termsSource2026-11-03
Hyperstack$2.50Not statedOn-demandNot statedNot statedSource2026-11-03
Hyperstack$2.60NVLOn-demandNot statedNot statedSource2026-11-03
Massed Compute$2.7380GBOn-demandNot statedSelf-serveSource2026-11-03
RunPod$2.89PCIeOn-demand30+ regions, not itemizedPods page lists Community Cloud and Secure Cloud tabs. The static page text does not say which tier the shown price is.Source2026-11-03
Civo$2.9980GBOn-demandNot statedSelf-serveSource2026-11-03
Massed Compute$3.14SXM5 80GBOn-demandNot statedSelf-serveSource2026-11-03
RunPod$3.19NVLOn-demand30+ regions, not itemizedPods page lists Community Cloud and Secure Cloud tabs. The static page text does not say which tier the shown price is.Source2026-11-03
Hyperstack$3.20SXMOn-demandNot statedNot statedSource2026-11-03
Thunder Compute$3.20SXMOn-demandNot statedPage: persistent GPU machines, ready to SSH into in minutesSource2026-11-03
Lambda$3.29PCIe 80GBOn-demandNot statedInstances: page says deploy in minutes, self-serve, first-come access. Reserved and clusters: contact sales.Source2026-11-03
Crusoe$3.9080GB HGXOn-demandNot statedPage: contact sales to lock in guaranteed resources at lowest ratesSource2026-11-03
Lambda$3.99HGX 80GBOn-demandNot statedInstances: page says deploy in minutes, self-serve, first-come access. Reserved and clusters: contact sales.Source2026-11-03
RunPod$3.99SXMOn-demand30+ regions, not itemizedPods page lists Community Cloud and Secure Cloud tabs. The static page text does not say which tier the shown price is.Source2026-11-03
Together AI$3.99HGXOn-demandNot statedNot statedSource2026-11-03
Lambda$4.09SXM 80GBOn-demandNot statedInstances: page says deploy in minutes, self-serve, first-come access. Reserved and clusters: contact sales.Source2026-11-03
Lambda$4.19SXM 80GBOn-demandNot statedInstances: page says deploy in minutes, self-serve, first-come access. Reserved and clusters: contact sales.Source2026-11-03
Lambda$4.29SXM 80GBOn-demandNot statedInstances: page says deploy in minutes, self-serve, first-come access. Reserved and clusters: contact sales.Source2026-11-03
DigitalOcean$4.41HGXOn-demandNot statedSelf-serveSource2026-11-03
Nebius$4.50HGXOn-demandNot statedReserved and committed: contact salesSource2026-11-03
Verda$4.84SXM5 80GBOn-demandNot statedReserved: from 1 month, custom quote (talk to sales)Source2026-11-03
CoreWeave$6.16HGXOn-demandNorth AmericaReserved and committed capacity: contact salesSource2026-11-03
Amazon Web Services$6.88SXM 80GB, 8xOn-demandUs-east-1Capacity Blocks for ML and quota limits apply; not shown in the price listSource2026-11-01
Amazon Web Services$6.8880GB, 1xOn-demandUs-east-1Capacity Blocks for ML and quota limits apply; not shown in the price listSource2026-11-01
Microsoft Azure$6.98NVL 94GB, 1x VMOn-demandEastusQuota and capacity approval apply; not shown by the price APISource2026-11-03
Oracle Cloud (OCI)$10.00Not statedOn-demandGlobal list priceNot statedSource2026-11-03
Microsoft Azure$12.29SXM 80GB, 8x VMOn-demandEastusQuota and capacity approval apply; not shown by the price APISource2026-11-03
Microsoft Azure$1.42NVL 94GB, 1x VMSpotEastusQuota and capacity approval apply; not shown by the price APISource2026-11-03
Hyperstack$2.00PCIeSpotNot statedNot statedSource2026-11-03
Microsoft Azure$2.27SXM 80GB, 8x VMSpotEastusQuota and capacity approval apply; not shown by the price APISource2026-11-03
Verda$2.42SXM5 80GBSpotNot statedNot statedSource2026-11-03
Nebius$0.79HGXPreemptibleNot statedNot statedSource2026-11-03
Together AI$1.99HGXPreemptibleNot statedNot statedSource2026-11-03
Modal$3.95SXM5ServerlessNot statedSelf-serveSource2026-11-03
Hyperstack$1.75Not statedReservedreservation, term not stated on pageNot statedContact salesSource2026-11-03
Hyperstack$1.82NVLReservedreservation, term not stated on pageNot statedContact salesSource2026-11-03
Civo$2.4980GBReserved36 monthsNot statedNot statedSource2026-11-03
Civo$2.5980GBReserved24 monthsNot statedNot statedSource2026-11-03
Civo$2.6980GBReserved12 monthsNot statedNot statedSource2026-11-03
Hyperstack$2.72SXMReservedreservation, term not stated on pageNot statedContact salesSource2026-11-03
Civo$2.7980GBReserved6 monthsNot statedNot statedSource2026-11-03
FlexiCloud (broker)$2.95HGX SXMReserved12 months or longer, floor rate on the longest termBy quotePage: available now; lead time per region is quoted in writing within about one working day; B300 allocation is the tightestSource2026-11-03
Together AI$3.19HGXReserved91 to 180 daysNot statedNot statedSource2026-11-03
DigitalOcean$3.26HGXReserved12 monthsNot statedPage: new pricing effective August 1, 2026. Reserved: contact sales.Source2026-11-03
Together AI$3.45HGXReserved31 to 90 daysNot statedNot statedSource2026-11-03
Together AI$3.69HGXReserved3 to 30 daysNot statedNot statedSource2026-11-03
Lambda$5.54Not statedReserved2 weeks to 1 yearNot stated1-Click Cluster: talk to salesSource2026-11-03
Lambda$5.85Not statedReserved2 weeks to 1 yearNot stated1-Click Cluster: talk to salesSource2026-11-03
Lambda$6.16Not statedReserved2 weeks to 1 yearNot stated1-Click Cluster: talk to salesSource2026-11-03
FluidstackNo public priceNo public pricing page found (/pricing returns 404)Not statedContact salesNot statedContact salesSource2026-11-01
Together AINo public priceContact usHGXContact sales181+ daysNot statedContact salesSource2026-11-03
Voltage ParkNo public priceContact for pricingHGXContact sales6+ monthsNot statedPage: Long-Term Reserve, 6+ month contracts, 32 to 8000+ H100 GPUsSource2026-11-03
Google CloudNo public pricePricing page is rendered by script; static fetch returns no ratesNot statedNot publicNot statedNot statedSource2026-11-01
HyperbolicNo public priceOn-demand, reserved and private cloud offered. No rate in the home page textNot statedNot publicNot statedFAQ: launch an on-demand GPU instance in minutes. Reserved: shorter-term commitment, usually under one yearSource2026-11-01
Paperspace (DigitalOcean)No public pricePage shows $3.09/hour on the H100 card and a footnote saying $5.95/hour on-demand promo for H100Not statedNot publicNot statedNot statedSource2026-11-01
SF ComputeNo public priceCalculator on site, no static rate in page textNot statedNot publicNot statedNot statedSource2026-11-01
Vast.aiNo public priceMarket-set. No static rate on the pricing pageNot statedNot publicNot statedSelf-serveSource2026-11-01
How to read this. Node prices are divided by the GPU count. A “from” price can be for a different variant, a smaller instance or a region that is out of stock. Compare all GPUs in the full price table.

FAQ

H100 pricing questions

Where is the cheapest place to rent an H100?

In our latest data, the lowest public on-demand H100 price is $1.68 per GPU-hour at Latitude.sh (80GB). Next lowest: Voltage Park at $1.99, Hyperstack at $2.50. A low headline price can depend on region, variant and stock, so open the source link before you buy.

How much does an H100 cost per hour to rent?

Across 28 public on-demand listings from 17 providers, the median H100 price is $3.99 per GPU-hour. Listings range from $1.68 to $12.29 per GPU-hour.

Is spot H100 pricing cheaper than on-demand?

The lowest spot or preemptible H100 price in our data is $0.79 per GPU-hour at Nebius, from 6 listings. The lowest on-demand price is $1.68. Spot capacity can be taken back at short notice, so use it for jobs that can restart.

How much does reserved H100 capacity cost?

The lowest published reserved H100 price in our data is $1.75 per GPU-hour at Hyperstack. Most reserved deals are quoted privately, so ask for the rate, term and deposit in writing.

How current are these H100 prices?

Every row shows the date it was read. The newest H100 rows were checked on Nov 3, 2026 (49 of 58 rows). Older rows keep their earlier date. Providers change prices without notice.

Need H100 capacity?

Tell us the count, term, region and budget. We reply with the public H100 prices that match and the checks to run before you pay. Free, and we do not sell compute.

Send a GPU request

Compare

Other GPU price guides

H200B200A100L40SB300GB200GB300MI300XAll GPU guidesFull price table