Hopper · data center GPU
Rent a dedicated NVIDIA H200 server for $1,690 a month
The H200 is the H100 with the memory it always wanted: the same Hopper SXM compute, 141 GB of HBM3e instead of 80 GB, and 4.8 TB/s of bandwidth instead of 3.35. For inference that means a 70B model at FP8 on a single GPU with 70 GB left for KV cache; for training it means larger micro-batches and fewer gradient-accumulation steps. Dedicated bare metal, one flat price a month.
- Per GPU, per month$1,690$2.32/h effective · fixed for your term
- Same GPU, hourly on-demand$3,270market median $4.48/h × 730 h
- You keep−48%$1,580 a month, every month
Specification
NVIDIA H200 server specifications
One dedicated machine, sized around the GPU. Nothing is shared, nothing is metered, and every line below is included in the monthly price.
Why teams rent the H200 here
- 141 GB HBM3e. 1.76× the memory of an H100 SXM on the same silicon — the cheapest way to fit a 70B model on one GPU.
- 4.8 TB/s bandwidth. 43% more than the H100. Decode-heavy inference is bandwidth-bound, and this is where the H200 pulls away.
- Drop-in for H100 code. Same Hopper architecture, same CUDA kernels, same FP8 Transformer Engine — nothing to port.
- One flat invoice. $1,690 a month covers the GPU, the host, the NVMe, 20 TB of transfer and 24/7 support — the same number every month, fixed for your term.
Below about 377 hours of use a month, an hourly provider is the cheaper way to run this GPU. Above it — and a machine that trains, serves or renders is above it — the monthly rate wins, and the gap is the $1,580 shown above.
Pricing
H200 price per month, by term
Per GPU, in USD, excluding VAT. Longer terms cost less; every term includes the same dedicated host, network and support.
| Term | Per GPU, per month | Effective hourly | vs on-demand | Order |
|---|---|---|---|---|
| MonthlyRolling month-to-month. Cancel with 30 days' notice. | $1,690 | $2.32/h | −48% | Deploy |
| 3 months5% off the monthly rate for a 3-month term. | $1,606 | $2.20/h | −51% | Deploy |
| 6 months10% off the monthly rate for a 6-month term. | $1,521 | $2.08/h | −53% | Deploy |
| 12 months15% off the monthly rate for a 12-month term. | $1,437 | $1.97/h | −56% | Deploy |
| 8× H200 nodeNVLink between the GPUs · NVLink 4 · 3.2 Tb/s InfiniBand · 192 vCPU · 2 TB RAM · 30 TB NVMe | $10,990 | $1,374/GPU | −19% | Cluster details |
Launch pricing: these rates apply to orders confirmed before Oct 31, 2026 and stay fixed for your whole term. From Nov 1, new H200 orders are billed at the list price of $1,949 a month. On-demand comparison: market-median published hourly rate for the same GPU ($4.48/h) over 730 hours. Block storage $25/TB/month and extra IPv4 addresses $4/month are optional add-ons.
Market comparison
H200 rental price compared
Every provider in our audit that publishes an on-demand rate for the H200, read on 19 September 2026 from their own pricing page and multiplied by 730 hours — what the same GPU costs kept for a month.
| Provider | Published rate, H200 | One month, 24/7 | vs our $1,690/mo |
|---|---|---|---|
 GPU Cloud HQDedicated, billed monthly |
$1,690 per month, flat | $1,690 | Our reference |
 CoreWeaveOn-demand, hourly |
$50.44/h per 8-GPU HGX H200 node — $6.31 per GPU-hour | $4,603 | −63% |
 Together AIOn-demand, hourly |
$5.99/h on demand · $2.99 preemptible | $4,373 | −61% |
 RunPodOn-demand, hourly |
$4.59/h Secure Cloud · $3.59 Community | $3,351 | −50% |
 NebiusOn-demand, hourly |
$4.50/h on demand · $2.45 preemptible | $3,285 | −49% |
 DigitalOcean (Paperspace)On-demand, hourly |
$4.47/GPU/h on demand · $3.40 with a 12-month commitment | $3,263 | −48% |
 Verda (DataCrunch)On-demand, hourly |
$4.37/h on demand · $2.19 spot | $3,190 | −47% |
 HyperstackOn-demand, hourly |
$3.99/h on demand · from $2.79 reserved | $2,913 | −42% |
| Market median30–40 providers, GetDeploying index | $4.48/h on demand | $3,270 | −48% |
Rates are each provider's standard on-demand tier for the H200 — not spot, preemptible or community hardware — as printed on their pricing page on 19 September 2026; each row links to the full comparison with its source. Where a provider is cheaper than us for a full month, the row says so. Method and caveats: GPU cloud alternatives.
Workloads
What people run on an H200
Large-model inference & training — and the three jobs below are where a dedicated H200 at a flat monthly price earns its keep.
-
LLM inference at scale
Llama-70B-class models on one GPU at FP8 with room for long prompts and big batches; 400B-class models on a single 8-GPU node.
-
Training with big batches
The extra memory goes into micro-batch size and sequence length — throughput you do not have to buy with more GPUs.
-
Multimodal and vision-language models
Image and video tokens are expensive in memory; 141 GB keeps encoder, decoder and KV cache on one card.
Is the H200 the right GPU for you?
Choose the H200 when an H100 workload is limited by memory or by decode bandwidth — inference above all. If the job is compute-bound and fits in 80 GB, the H100 SXM does the same work for less; if you need Blackwell-class throughput, the B200 is the step up.
H200 rental: questions answered
How much does it cost to rent an H200 server?
$1,690 a month, flat, for a dedicated NVIDIA H200 server with 24 vCPU · 256 GB RAM · 3 TB NVMe and a 25 Gbps uplink — $2.32 per hour effective over 730 hours. A 3-, 6- or 12-month term takes 5%, 10% or 15% off: $1,606, $1,521 or $1,437 a month. The market-median on-demand rate for the same GPU is $4.48 per hour, about $3,270 for a full month.
Is the H200 in stock right now?
Yes. 9 of the current batch of 32 were unallocated at the last stock update (18 September 2026). Provisioning is automated: your server is imaged, secured with your SSH key and handed over in under 10 minutes after your payment confirms. The next batch is due Oct 13.
What is included with a H200 server?
Everything on the pricing table: the GPU, 24 vCPU · 256 GB RAM · 3 TB NVMe, a 25 Gbps uplink with 20 TB of outbound transfer, a dedicated public IPv4 and a /64 IPv6 block, DDoS protection, root access and Ubuntu 24.04 with the NVIDIA driver, CUDA 12 and Docker — or PyTorch, vLLM, Kubernetes images at checkout. Hardware replacement within 4 hours and a 99.9% uptime SLA are part of the contract.
H200 or H100: which should I rent?
For training a model that fits comfortably in 80 GB, the H100 SXM is cheaper for identical compute. For inference, long contexts or anything that spills over 80 GB, the H200 wins: same architecture, 1.76× the memory, 43% more bandwidth, and one GPU where you would otherwise need two.
Can I rent several H200s or a full 8-GPU node?
Yes: order 1, 2, 4 or 8 GPUs at once, subject to the units left in the batch, on the same monthly terms. The 8× H200 node — NVLink between the GPUs, 3.2 Tb/s InfiniBand, NVLink 4 · 3.2 Tb/s InfiniBand · 192 vCPU · 2 TB RAM · 30 TB NVMe — is priced separately at $10,990 a month on the clusters page.
Do I need KYC or a card to rent it?
No. An email address is all we ask for. You fund a prepaid balance in Bitcoin, Ether or USDT (ERC-20 or TRC-20); the first invoice settles from it and monthly renewals charge to it automatically. No ID document, no card, no bank transfer.
Other data center GPUs by the month
Same dedicated hosts, same monthly model, same 10-minute provisioning. All 10 GPUs are on the pricing page.
Deploy an H200 in under 10 minutes
$1,690 a month, fixed for your term. Fund your balance in BTC, ETH or USDT and the server is imaged, secured with your key and handed over automatically.