Loading prices…
Lowest on-demand
$1.66per GPU·h · ≈ $1,212/month
at Latitude.shMedian on-demand
$3.22/GPU·h
across 12 providers
Lowest spot
$2.00/GPU·h
Reserved from $1.18
FP16/BF16 cost
$2.20/PFLOP·h
#21 of 69 GPUs for value
Providers
13
39 configurations
Price history
Daily H100 PCIe prices per GPU-hour. The shaded period is reconstructed from public archives: the version history of SkyPilot's price catalogs, Internet Archive copies of providers' pricing pages and the open gpu-price-feed project (fewer providers than live data). After that, PetaPrice's own tracking.
Lowest on-demandMedian on-demandLowest spot
Price by provider
Cheapest On-demand H100 PCIe offer at each provider, per GPU-hour. Shorter and greener is cheaper.
Latitude.sh🇺🇸$1.66best
RunPod🇺🇸$1.99+20%
IMWT$2.48+49%
Hyperstack🇬🇧$2.50+51%
Massed Compute🇺🇸$2.73+64%
OVHcloud🇫🇷$3.13+89%
Scaleway🇫🇷$3.30+99%
Oblivus🇬🇧$3.502.1×
Sesterce🇫🇷$3.632.2×
Lambda🇺🇸$3.792.3×
T Cloud Public🇩🇪$3.822.3×
Paperspace🇺🇸$5.993.6×
Thin vertical line = median across providers ($3.13).
Every provider we track
Pricing
Size
Host
Provider HQ
13 providers · 39 configurations
Sort
- Latitude.shOn-demandReserved🇺🇸 HQ in United Statesvia Shadeform, Latitude pricing pagefrom · On-demandCheapest$1.66/GPU·hCheapest
- RunPodcommunityOn-demand🇺🇸 HQ in United Statesvia RunPodfrom · On-demand+20%$1.99/GPU·h+20%
- IMWTOn-demandHQ unknownvia Shadeformfrom · On-demand+49%$2.48/GPU·h+49%
- 8× GPU · PCIe128 vCPU · 512 GB RAM · 5,000 GB disk$2.48 /GPU·h$19.8 /hOn-demand🇺🇸 US, Des Moines, IA○ sold outRent ↗
- 2× GPU · PCIe40 vCPU · 256 GB RAM · 2,500 GB disk$2.48 /GPU·h$4.96 /hOn-demand🇺🇸 US, Des Moines, IA○ sold outRent ↗
- 4× GPU · PCIe64 vCPU · 512 GB RAM · 5,000 GB disk$2.48 /GPU·h$9.92 /hOn-demand🇺🇸 US, Des Moines, IA○ sold outRent ↗
- HyperstackOn-demandReservedSpot🇬🇧 HQ in United Kingdomvia Hyperstack pricing pagefrom · On-demand+51%$2.50/GPU·h+51%
- Massed ComputeOn-demand🇺🇸 HQ in United Statesvia Massed Compute pricing pagefrom · On-demand+64%$2.73/GPU·h+64%
- OVHcloudOn-demand🇫🇷 HQ in Francevia OVHcloud catalog APIfrom · On-demand+89%$3.13/GPU·h+89%
- ScalewayOn-demand🇫🇷 HQ in Francevia Shadeformfrom · On-demand+99%$3.30/GPU·h+99%
- OblivusOn-demandReserved🇬🇧 HQ in United Kingdomvia Oblivus pricing pagefrom · On-demand2.1×$3.50/GPU·h2.1×
What can the H100 PCIe run?
Memory for the weights plus one 32K-token sequence of context, how many GPUs hold it, and what that replica costs at the cheapest in-stock on-demand rate.
| Model | Params | Memory | GPUs | $/hr | $/mo |
|---|---|---|---|---|---|
gpt-oss-20b OpenAI · 20.9B | 20.9B · 3.6B active | 20 GiB | 1 | $1.99 at RunPod | $1,453 |
Gemma 3 27B Google · 27.4B | 27.4B | 28 GiB | 1 | $1.99 at RunPod | $1,453 |
Qwen3-32B Alibaba · 32.8B | 32.8B | 39 GiB | 1 | $1.99 at RunPod | $1,453 |
Llama 3.3 70B Instruct Meta · 70.6B | 70.6B | 76 GiB | 2 | $3.98 at RunPod | $2,905 |
gpt-oss-120b OpenAI · 117B | 117B · 5.1B active | 110 GiB | 2 | $3.98 at RunPod | $2,905 |
Qwen3-235B-A22B Alibaba · 235B | 235B · 22B active | 225 GiB | 4 | $7.96 at RunPod | $5,811 |
DeepSeek-R1 DeepSeek · 684B | 684B · 37B active | 640 GiB | 16 | $31.84 at RunPod | $23,243 |
Kimi K3 Moonshot AI · 2.78T | 2.78T · 104B active | 2.53 TiB | 64 | $127.36 at RunPod | $92,973 |
Assumes 90% of each GPU's memory is usable and rounds up to 1, 2, 4, 8… GPUs for tensor parallelism (up to 64). $/hr is the GPU count times the cheapest in-stock on-demand rate per GPU; check that the provider rents instances that large. Serving many requests at once needs more memory for context.
Alternatives to the H100 PCIe
GPUs with similar compute and memory, and their lowest on-demand price right now.
H100 PCIe pricing questions
How much does it cost to rent an H100 PCIe?
On-demand H100 PCIe prices start at $1.66 per GPU-hour (Latitude.sh). The median across 12 providers is $3.22, and the most expensive is $5.99. Running one GPU around the clock costs about $1,212 per month at the lowest price.
Is there cheaper spot or interruptible H100 PCIe capacity?
Yes. Spot (interruptible) H100 PCIe capacity starts at $2.00 per GPU-hour, -20% below the cheapest on-demand price. Spot instances can be preempted at any time, so use them for checkpointed training or batch jobs.
How much compute do you get per dollar with the H100 PCIe?
At $1.66 per hour and 0.756 PF of dense FP16/BF16, the H100 PCIe costs $2.20 per PFLOP-hour. That ranks 21 of 69 GPUs with on-demand prices on PetaPrice. Figures are datasheet peaks; real utilization is typically 30–60%.
What are the H100 PCIe's key specs?
80 GB HBM2e with 2,000 GB/s of memory bandwidth, Hopper architecture (2022), PCIe form factor, 350 W. Dense peak: 0.756 PF FP16 / BF16, 1.51 PF FP8, 1.51 PF INT8, 0.378 PF TF32, 0.051 PF FP32, 0.051 PF FP64.
How current are these prices?
Prices were collected on Nov 9, 2026, 21:02 UTC from provider APIs and public catalogs and refresh about every 2 hours. They are indicative list prices; always confirm on the provider's website before booking.