GB300 NVL72 puts 72 Blackwell Ultra GPUs and 36 Grace CPUs in a single 130TB/s NVLink domain, with 288GB of HBM3e per GPU. For training, that keeps tensor and expert-parallel traffic inside the rack instead of crossing slower node-to-node networking, which matters most for frontier-scale pretraining and very large mixture-of-experts models. It is the same silicon as B300 and the Blackwell Ultra refresh of GB200's rack design, with more memory per GPU at higher power draw. Tracked on-demand GB300 price per hour spans roughly $4.00 to $18.00 per GPU, and most volume sits in reserved terms. Full GB300 NVL72 specs are available on request. B300, B200 and H200 for LLM training remain the more available options for runs that fit on a few nodes.
GB300 NVL72 is Blackwell Ultra as a full rack: 72 GPUs, 36 Grace CPUs, 288GB of HBM3e per GPU and 20TB in aggregate, all in one NVLink domain. For training, the value is communication: parallelism strategies that would normally cross node boundaries stay on NVLink, which is where very large pretraining runs and mixture-of-experts models lose the most efficiency on smaller clusters. It is the same silicon as B300, and the same 72-GPU layout as GB200 with more memory and compute per GPU, at roughly 1,400W per GPU and 135-140kW per rack. For runs that fit comfortably on a few B300, B200 or H200 nodes, the rack adds cost without a matching gain, and its supply sits with operators who control their own power and cooling, so it is a deliberate choice for frontier-scale training rather than a default upgrade.
GB300 training is placement-sensitive: rack-scale power and liquid cooling make confirming real in-country supply and lead times matter more than for any single-card generation. See GB300 availability by country below.
Read the full guide to GPU cloud in this location →Every location links to its own page. Click through for local pricing and specs.
Wholesale rates against cloud list price for a 64-GPU cluster.
We connect you to our vetted partners. You contract directly with the operator running your nodes.
GPU model, count, placement and timeline. Add workload detail if you have it.
We find vetted partners with capacity that fits, in the jurisdiction you need.
Real quotes from partners who hold the capacity, not listings that may not exist.
You contract directly with the operator. We smooth the provisioning process.
—
Need a different GPU generation? Each model available here has a dedicated page with full pricing and specifications.
Tell us the essentials. We'll line up real quotes from our vetted wholesale partners, and you contract directly with the operator.
GPUaaS.com connects enterprises with vetted GPU providers for large-scale compute at up to 30% lower cost than hyperscalers.
