Skip to content
aipowered

Find your next machine.

Try “5090”, “Jetson”, or “DGX”.

NVIDIA GB200 (186GB HBM3e)

186GB HBM3e
Memory
8,000GB/s
Memory bandwidth
Q4 2024
Released

Price history

On-demandSpot
USD / GPU / HR
On-demand · latest$9.992026-10-05

What it costs to rent an NVIDIA GB200

The NVIDIA GB200 is a Blackwell-architecture training NVIDIA GPU with 186GB of HBM3e memory and 8,000 GB/s of memory bandwidth, released Q4 2024. It is built for training frontier-scale models and high-throughput production inference.

Renting one currently costs between $4.99 and $16.00 per GPU-hour across the 3 cloud providers we track, with a median of $9.99. The cheapest live rate is $4.99/hr from Together (On-demand). All rates below are normalised to a single GPU per hour so providers with multi-GPU nodes compare like-for-like.

Where to rent the GB200

How we rank

The median is one provider, one vote: each provider contributes the median of its own on-demand listings, then the headline is the median across providers — so no single marketplace or hyperscaler sets the price alone. Below two providers we show the range instead.

Default order is relevance, weighed the way a buyer would: price per GPU first, lightly blended with total node cost; on-demand preferred over spot and reserved when prices are close; established clouds edged above marketplaces; in-stock, recently-checked offers above stale ones; and no single provider is allowed to crowd the top. Click a column to sort by the raw numbers. No paid placements — ranking is never sold.

Scroll horizontally to compare all columns →

ProviderTierConfigTotal VRAMBillingAvailabilityGo to provider
TogetherBestEnterprise1× GB200186GBOn-demand$4.99$4.99Rent
Fal.aiGPU cloud1× GB200186GBOn-demand$9.99$9.99Rent
Oracle CloudHyperscaler1× GB200186GBOn-demand$16.00$16.00Rent

Technical specifications

Throughput
FP6445 TFLOPS
FP3290 TFLOPS
TF321,250 TFLOPS
FP16/BF162,500 TFLOPS
FP85,000 TFLOPS
FP410,000 TFLOPS
Silicon & memory
Memory
13,824 GB
Memory bandwidth
8,000 GB/s
TDP
120,000 W
INT8
5,000 TOPS
Foundry
TSMC
Released
15 Feb 2025
TrainingInferenceAI Factory
Datasheet
GPUs72x NVIDIA B200 (Blackwell)
CPUs36x NVIDIA Grace
FP4 compute1.44 EFLOPS (sparse)
HBM3eabout 13.5 TB
NVLink domain130 TB/s
Coolingliquid, full rack
Deployment120 kW class per rack
Class
Training
Cloud providers
3 tracked
Price range
$4.99 – $16.00 / GPU / hr

SNAPSHOT 2026-10-12T05:40:41.151564+00:00 · BACK TO THE FULL INDEX