Ampere · data center GPU
Rent a dedicated NVIDIA A100 server for $649 a month
The A100 80 GB is still the best-value data center GPU for a large class of work: 80 GB of HBM2e at 2 TB/s, 312 TFLOPS of FP16 tensor compute, TF32 for classic deep learning and NVLink 3 at 600 GB/s. It has no FP8 — that arrived with Hopper — but for fine-tuning, computer vision, speech and 16-bit inference it does the same job as an H100 for about half the monthly price. Dedicated bare metal, 16 vCPU, 128 GB of RAM and 2 TB of NVMe included.
- Per GPU, per month$649$0.89/h effective · fixed for your term
- Same GPU, hourly on-demand$1,285market median $1.76/h × 730 h
- You keep−49%$636 a month, every month
Specification
NVIDIA A100 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 A100 here
- The cheapest 80 GB of HBM in the catalogue. The same memory capacity as an H100 SXM at about half the monthly rate — and memory per dollar is what most fine-tunes are bound by.
- Mature, supported everywhere. Every framework, kernel and container image has been tuned on Ampere for years; nothing surprises you.
- MIG partitioning. Split one A100 into up to seven isolated instances to run many small inference services on one GPU.
- One flat invoice. $649 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 369 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 $636 shown above.
Pricing
A100 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. | $649 | $0.89/h | −49% | Deploy |
| 3 months5% off the monthly rate for a 3-month term. | $617 | $0.85/h | −52% | Deploy |
| 6 months10% off the monthly rate for a 6-month term. | $584 | $0.80/h | −55% | Deploy |
| 12 months15% off the monthly rate for a 12-month term. | $552 | $0.76/h | −57% | Deploy |
On-demand comparison: market-median published hourly rate for the same GPU ($1.76/h) over 730 hours. Block storage $25/TB/month and extra IPv4 addresses $4/month are optional add-ons.
Market comparison
A100 rental price compared
Every provider in our audit that publishes a price for the A100, read on 15 October 2026 from their own pricing page: hourly rates multiplied by 730 hours — what the same GPU costs kept for a month — and monthly prices as listed.
| Provider | Published rate, A100 | One month, 24/7 | vs our $649/mo |
|---|---|---|---|
 GPU Cloud HQDedicated, billed monthly |
$649 per month, flat | $649 | Our reference |
 Google CloudOn-demand, hourly |
$5.07/h for a2-ultragpu-1g (1× A100 80 GB) | $3,701 | −82% |
 Microsoft AzureOn-demand, hourly |
$3.67/h for NC24ads A100 v4 (1× A100 80 GB PCIe) | $2,681 | −76% |
 AWS (Amazon EC2)On-demand, hourly |
$27.45/h for p4de.24xlarge (8× A100 80 GB) — $3.43 per GPU-hour | $2,504 | −74% |
 LambdaOn-demand, hourly |
$2.79/h (SXM 80 GB, 8-GPU instances) · $1.99 for 40 GB | $2,037 | −68% |
 CoreWeaveOn-demand, hourly |
$21.60/h per 8-GPU A100 node — $2.70 per GPU-hour | $1,971 | −67% |
 VultrOn-demand, hourly |
$2.397/h for 1× A100 80 GB PCIe · $2.80/GPU/h for an 8× HGX A100 | $1,750 | −63% |
 Crusoe CloudOn-demand, hourly |
$2.30/GPU-h (SXM 80 GB) · $2.00 PCIe | $1,679 | −61% |
 Cherry ServersDedicated, monthly |
from €1,416.82/mo ($1,658) — A100 80 GB in a custom bare-metal server, waiting list | $1,658 | −61% |
 Verda (DataCrunch)On-demand, hourly |
$2.00/h on demand (SXM4 80 GB) · $1.00 spot | $1,460 | −56% |
 HOSTKEYDedicated, monthly |
€1,300/mo ($1,450) — 1× A100 80 GB virtual server, pre-order | $1,450 | −55% |
 TensorDockOn-demand, hourly |
from $1.80/h (SXM4) | $1,314 | −51% |
 RunPodOn-demand, hourly |
$1.79/h Secure Cloud (SXM) · $1.39 Community | $1,307 | −50% |
 OVHcloudCloud, monthly billing |
$1,229/mo for a100-180 (1× A100 80 GB) · $3.07/h by the hour | $1,229 | −47% |
 HyperstackOn-demand, hourly |
$1.60/h SXM on demand · from $1.36 reserved | $1,168 | −44% |
 JarvislabsOn-demand, hourly |
$1.49/h on demand (80 GB) · $0.89 spot | $1,088 | −40% |
 Massed ComputeOn-demand, hourly |
$1.38/h (SXM4 80 GB) | $1,007 | −36% |
 Thunder ComputeOn-demand, hourly |
$1.09/h at 1–2 GPUs · $1.49 per GPU at 4–8 | $796 | −18% |
 Vast.aiOn-demand, hourly |
median $0.51/h, SXM4 (14 Oct) · offers from $0.40 | $372 | +74% |
| Market median30–40 providers, GetDeploying index | $1.76/h on demand | $1,285 | −49% |
Rates are each provider's standard on-demand tier for the A100 — not spot, preemptible or community hardware — or its listed monthly price where it publishes one, as printed on their pricing page on 15 October 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; every GPU at a glance: GPU cloud pricing.
Workloads
What people run on an A100
Proven workhorse for training — and the three jobs below are where a dedicated A100 at a flat monthly price earns its keep.
-
Fine-tuning and RLHF
LoRA and QLoRA on 70B models, full fine-tunes up to 13B across two to four cards, DPO and reward-model training — 80 GB per GPU keeps the optimizer state on the cards.
-
Computer vision, speech and classic deep learning
Detection, segmentation, ASR and recommender models that live in TF32/FP16 and never needed FP8.
-
16-bit inference
Models up to about 35B at FP16 on one GPU, or many small models side by side with MIG.
Is the A100 the right GPU for you?
Choose the A100 when the workload runs in FP16/BF16 or TF32 and the budget matters: fine-tuning, CV, speech, embeddings. Move to the H100 SXM when you can use FP8 or need the extra compute for a long training run, and to the L40S when 48 GB is enough and the job is inference or graphics.
Recommended for fine-tuning & rlhf — the sizing guide there shows what fits in 80 GB HBM2e.
Compare the A100 side by side with the cards people weigh it against — specifications, price per GB and the model sizes that fit — or check what your own model needs with the GPU memory calculator.
At FP8 with an 8k context, one A100 runs these open models on its own — each page has the memory by context length and the fine-tuning needs:
A100 rental: questions answered
How much does it cost to rent an A100 server?
$649 a month, flat, for a dedicated NVIDIA A100 server with 16 vCPU · 128 GB RAM · 2 TB NVMe and a 25 Gbps uplink — $0.89 per hour effective over 730 hours. A 3-, 6- or 12-month term takes 5%, 10% or 15% off: $617, $584 or $552 a month. The market-median on-demand rate for the same GPU is $1.76 per hour, about $1,285 for a full month.
Is the A100 in stock right now?
Yes. 21 of the current batch of 48 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.
What is included with a A100 server?
Everything on the pricing table: the GPU, 16 vCPU · 128 GB RAM · 2 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.
Is the A100 still worth renting in 2026?
For fine-tuning, computer vision, speech and 16-bit inference, yes — it delivers the same 80 GB as an H100 for about half the monthly price, and every framework is tuned for it. It is the wrong choice for FP8 training or for very long LLM training runs, where the H100 SXM finishes sooner and costs less per epoch.
Can I rent several A100s?
Yes: order 1, 2, 4 or 8 GPUs at once, subject to the units left in the batch, on the same monthly terms.
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 A100 in under 10 minutes
$649 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.