NVIDIA B200 / BLACKWELL / 8 GPUS PER SERVER
NVIDIA B200 cluster rental price: on-demand and reserved
Published on-demand rates for a B200 run from $5.43 to $14.24 per GPU-hour. Reserved clusters cost less. The lowest published 36-month rate is $3.79. More providers publish B200 terms than B300 terms, so the reserved range is clearer here.
Updated October 17, 2026. By GPU Rentals, powered by Bridgepointe Technologies. Every figure comes from the provider's own pricing page. Sources are listed at the end.
Per GPU-hour, 11 providers
Per GPU-hour, 3 providers
Per GPU-hour, 2 providers
Our B200 reserved cluster rates
We compare against the lowest of AWS, Microsoft and Google for each term. Rates are per GPU-hour, billed monthly, with no upfront payment.
| Term | GPU Rentals rate per GPU-hour | Lowest AWS, Microsoft or Google rate |
|---|---|---|
| 12 months | $6.96 | $11.12, Google Cloud 1-year commitment |
| 24 months | $5.41 | Not offered |
| 36 months | $3.85 | $6.15, AWS 3-year Savings Plan |
| 60 months | $3.47 | Not offered |
AWS, Microsoft and Google sell only 1-year and 3-year terms, so 24 and 60 months have no direct comparison. Every rate depends on confirmed capacity for your GPU count, start date and region.
B200 on-demand price per GPU-hour
On-demand means you pay by the hour and can stop anytime. Most providers rent B200 in whole 8-GPU servers.
| Provider | On-demand per GPU-hour | How it is sold |
|---|---|---|
| Massed Compute | $5.43 | 8-GPU instance at $43.46 per hour |
| Runpod | $5.98 to $7.99 | Depends on the cloud tier |
| Hyperstack | $6.00 | Per GPU |
| Lambda | $6.69 to $6.99 | $6.69 on an 8-GPU instance |
| Packet.ai | $6.99 | Dedicated GPU |
| Verda | $7.99 | Per GPU |
| Together AI | $8.19 | GPU cluster |
| Nebius | $8.50 | Per GPU |
| CoreWeave | $8.60 | 8-GPU node at $68.80 per hour |
| Oracle Cloud | $14.00 | BM.GPU.B200.8 bare metal |
| AWS | $14.24 | p6-b200.48xlarge, 8 GPUs, N. Virginia |
Spot and preemptible rates are left out. They are cheaper, but the provider can take the GPUs back.
B200 reserved cluster pricing by term
A reserved term commits you to a set number of GPUs for months or years. In return the hourly rate drops. Longer terms drop it further.
| Provider | 12 months | 24 months | 36 months | Notes |
|---|---|---|---|---|
| Civo | $4.49 | $3.99 | $3.79 | 8-GPU instance. |
| Verda | $7.35 | $5.99 | Not published | Our math: published 8% and 25% term discounts applied to $7.99. |
| Google Cloud | $11.12 | Not offered | $7.09 | A4 VM, 8 GPUs, committed use discount. |
Short terms
Some providers publish terms of days or months, between on-demand and a full year. Together AI charges $7.99 for 3 to 30 days, $7.79 for 31 to 90 days and $6.79 for 91 to 180 days. AWS sells prepaid Capacity Blocks of 1 to 182 days at $14.21.
Cluster size changes the price
Lambda publishes B200 cluster rates by size for terms of two weeks to one year:
| Cluster size | Per GPU-hour |
|---|---|
| 16 GPUs | $9.86 |
| 64 GPUs | $9.36 |
| 256 GPUs or more | $8.87 |
Nebius, CoreWeave, Runpod, Crusoe, Massed Compute and Microsoft Azure quote reserved B200 privately. Hyperstack says reserved starts "from $5.10" and Vultr says prepaid starts "from $3.50", neither with a term. Nebius says reserved runs "up to 35% less" than on-demand.
How B200 clusters are sized: 8, 64, 512 GPUs
A B200 is sold inside an 8-GPU server, so cluster sizes move in steps of 8. A request for 500 GPUs becomes 496 or 504 in a real quote. Training clusters usually grow in powers of two: 64, 128, 256, 512 GPUs.
The B200 is one generation behind the B300. It has less memory per GPU and costs less per hour. For models that fit comfortably, a B200 cluster is often the better value. For the largest models, look at the GB300 NVL72.
Large clusters rarely come through self-serve checkout. New cloud accounts usually start with a quota of zero high-end GPUs, and big orders go through a sales contract.
GET A B200 CLUSTER QUOTE
Reserved B200 clusters.
Checked for you.
Tell us how many B200s you need, for how long, and when. We check providers and reply by email with options and pricing. Nothing is booked or charged.
REQUEST SENT
Thanks. We have your B200 request.
We will reply by email with provider options and pricing. Nothing has been booked or charged.
B200 CLUSTER QUESTIONS
Before you rent a B200 cluster.
What are GPU Rentals' B200 cluster rates?
Our reserved B200 rates are $6.96 per GPU-hour for 12 months, $5.41 for 24 months, $3.85 for 36 months and $3.47 for 60 months, billed monthly. Rates are subject to confirmed capacity.
How much does it cost to rent a B200?
Published on-demand rates run from $5.43 to $14.24 per GPU-hour. A full 8-GPU server at $6.69 per GPU-hour costs about $38,500 for a 30-day month.
How much does a 64-GPU B200 cluster cost per month?
At published on-demand rates, 64 GPUs cost about $250,000 to $656,000 for a 30-day month. At the lowest published 36-month reserved rate of $3.79, the same cluster is about $175,000 a month.
How much does a 512-GPU B200 cluster cost per month?
At published on-demand rates, 512 GPUs cost about $2.0 million to $5.25 million for a 30-day month. At $3.79 on a 36-month term, it is about $1.4 million a month. Clusters this size are negotiated, so a quote can come in lower.
Is a reserved B200 cluster cheaper than on-demand?
Yes. Published 12-month rates run from $4.49 to $11.12 and 36-month rates from $3.79 to $7.09 per GPU-hour. Larger clusters also get lower rates. Lambda charges $9.86 per GPU-hour for 16 GPUs and $8.87 for 256 or more.
Should I rent B200 or B300?
The B300 is newer, with 288 GB of memory per GPU. The B200 costs less per hour. If your model fits on B200, a B200 cluster usually gives more compute per dollar.
How many B200s come in one server?
Eight. Providers rent B200 in whole 8-GPU servers, so cluster sizes go up in steps of 8.
Sources, checked October 17, 2026
Massed Compute: vm.massedcompute.com/pricing. Runpod: runpod.io/pricing, updated October 8, 2026. Hyperstack: hyperstack.cloud/gpu-pricing. Lambda: lambda.ai/pricing. Packet.ai: packet.ai/pricing. Verda: verda.com/pricing. Together AI: together.ai/pricing. Nebius: nebius.com/prices. CoreWeave: coreweave.com/pricing. Oracle: oracle.com/cloud/price-list. AWS on-demand price file, October 15, 2026: aws.amazon.com/ec2/pricing/on-demand. AWS Capacity Blocks: aws.amazon.com/ec2/capacityblocks/pricing. Civo: civo.com/pricing; it lists three term prices for its 8-GPU instance, read here as 12, 24 and 36 months in page order. Google Cloud: cloud.google.com. Vultr: vultr.com/pricing.
Prices change often and are shown as published. Monthly figures use 720 hours. A published price is not a promise that capacity is available.