NVIDIA Blackwell Ultra

NVIDIA B300 GPU cloud

Reserve NVIDIA B300 (Blackwell Ultra) capacity on VESSL Cloud: up to 288GB HBM3e and FP4 acceleration for the largest models and high-concurrency reasoning inference.

Reserve capacity See pricing & availability

NVIDIA B300

GPU memory: up to 288GB HBM3e
Memory bandwidth: 8 TB/s

Technical specifications

Architecture: Blackwell
GPU memory: up to 288GB HBM3e
Memory bandwidth: 8 TB/s
NVLink: 1.8 TB/s
FP8 (Tensor): 9 PFLOPS
FP4 (Tensor): 18 PFLOPS
Max TDP: 1,100W
GPUs per node: 8 (HGX B300)

*Peak performance with sparsity, per NVIDIA official specs. Final specs may vary by node configuration.

Pricing & availability

NVIDIA B300Reserved only

Talk to sales

What's the B300 best for?

Largest-model training

Up to 288GB HBM3e per GPU and 1.8 TB/s NVLink keep trillion-parameter models resident with fewer partitions and less communication overhead.

Reasoning & long-context inference

Blackwell Ultra's huge HBM3e holds massive KV caches, and roughly 1.5× the B200's dense FP4 compute serves agentic and reasoning workloads at high concurrency.

Consolidate inference fleets

With up to 288GB HBM3e per GPU, a large model can fit on fewer B300s than on 192GB B200s.

Compare NVIDIA data-center GPUs

	A100 Ampere	H100 Hopper	H200 Hopper	B200 Blackwell	B300 You're viewing	GB300 Blackwell
Architecture	Ampere	Hopper	Hopper	Blackwell	Blackwell	Blackwell
GPU memory	80GB HBM2e	80GB HBM3	141GB HBM3e	192GB HBM3e	up to 288GB HBM3e	up to 288GB HBM3e
Memory bandwidth	2.0 TB/s	3.35 TB/s	4.8 TB/s	8 TB/s	8 TB/s	8 TB/s
FP8 (Tensor)	Not supported	3,958 TFLOPS	3,958 TFLOPS	9 PFLOPS	9 PFLOPS	10 PFLOPS
Access	from $1.48/hr	from $2.98/hr	Reserved only	Reserved only	Reserved only	Reserved only
Best for	Cost-efficient fine-tuning & inference	Cost-efficient training & inference	Long-context & large-model inference	Frontier-scale training (FP4)	Largest models & reasoning inference	Cluster-scale training & inference

Why AI teams run GPUs on VESSL Cloud

From A100 to Blackwell Ultra

Run A100 and H100 on demand. Reserve H200 and Blackwell GPUs for larger runs, so each workload gets the GPU it needs.

Issues caught early

Every VM Cluster node is health-checked continuously, with its status refreshed about once a minute. Workspaces and Jobs alert you when a run crosses a resource threshold or hits a runtime issue.

From one node to a cluster

Train and serve on up to 8 GPUs on one node, and move to a multi-node VM Cluster when one node is not enough.

Enterprise-ready

SOC 2 Type II compliance, with dedicated support for production AI.

Frequently asked questions

How do I get access to NVIDIA B300 GPUs?

B300 (Blackwell Ultra) capacity is allocated on request. Talk to our team for current availability and pricing.

How much memory does the B300 have?

The B300 (Blackwell Ultra) has up to 288GB HBM3e per GPU. Talk to our team for current node configurations and availability.

What's the difference between the B200 and B300?

The B300 (Blackwell Ultra) increases memory to up to 288GB HBM3e (vs 192GB on the B200) and delivers roughly 1.5× the B200's dense FP4 compute, built for the largest models and high-concurrency reasoning inference.

Is the B300 better for training or inference?

Both. FP4/FP8 acceleration and up to 288GB HBM3e make the B300 ideal for frontier-scale training and high-throughput, low-latency reasoning inference.

Can I reserve a full B300 cluster?

Yes. We provision HGX B300 nodes (8 GPUs each), from a single node to multi-node clusters, joined by InfiniBand or RoCE, depending on the cluster. How many nodes you can reserve depends on current supply, so talk to our team for availability.