Log in Rent B300
B300 288GB · live marketplace — from $4.77 / hr

Rent a B300
by the minute.
Not the month.

Rent a B300, the Blackwell Ultra ceiling. 288 GB HBM3e at 8,000 GB/s, FP4 tensor cores, 5th-gen NVLink at 1.8 TB/s. The extra 96 GB over B200 unlocks 1T+ parameter pretraining without aggressive sharding, DeepSeek-V3 671B MoE serves at ~5,200 tok/s aggregated with full KV cache headroom, Llama-3 70B FP4 at ~7,500 tok/s with 256-request concurrency. Available on Clore.ai bare-metal 8-GPU NVLink-Switch pods. Billed per-minute, paid in BTC or USDT/USDC.

Per-minute billing SSH + Docker + Jupyter Spot & on-demand Texas & EU regions
$4.77/hr
Starting spot price
288GB
HBM3e VRAM per card
8TB/s
Memory bandwidth per card
<90s
Cold-start to ready
workloads

Trillion-parameter silicon,
by the minute.

B300s are what you rent when memory matters. 288 GB HBM3e per card holds 1T+ MoE active weights with margin. FP4 tensor cores, 8 TB/s HBM keeps the SMs saturated; 5th-gen NVLink lets eight cards train as one.

288 GB HBM3e — trillion-parameter inference

B300 raises the per-card memory ceiling to 288 GB HBM3e, 50% more than B200. This is the unlock for 1T+ parameter pretraining and serving 671B MoE with full KV cache resident on one node. Blackwell Ultra tensor cores keep B200's FP4/FP8 throughput while giving you the memory headroom that B200 couldn't sustain.

1T+ MoE pretrain ~16k tok/s/GPU

Production LLM serving

TensorRT-LLM, vLLM, SGLang — all tuned for Blackwell Ultra. 288 GB VRAM serves 70B FP4 with massive KV-cache headroom, or DeepSeek-V3 671B MoE from an 8-GPU pod.

Llama-3 70B FP4 ~7,500 tok/s

Long-context inference

288 GB HBM3e + FlashAttention-3 means 256k-token contexts run without offload, and 671B MoE fits with full KV cache resident on a single 8-GPU pod. Ideal for long-doc RAG and agent loops with tool use.

256k ctx, batch-8 no offload
why B300

The largest-VRAM GPU
you can rent by the minute.

Blackwell Ultra is the architecture trillion-parameter foundation models are being pretrained on. Specs from Nvidia's SXM5 datasheet; pricing reflects the lowest live spot floor.

B300 288GB B200 192GB H200 141GB H100 80GB
Architecture Blackwell Ultra Blackwell Hopper Hopper
VRAM 288 GB HBM3e 192 GB HBM3e 141 GB HBM3e 80 GB HBM3
Memory bandwidth 8,000 GB/s 8,000 GB/s 4,800 GB/s 3,350 GB/s
TDP 1,400 W 1,000 W 700 W 700 W
NVLink fabric 1.8 TB/s 5th-gen 1.8 TB/s 5th-gen 900 GB/s 900 GB/s
From / hr (spot) $4.77 $3.40 $2.40 $1.89

// prices are spot-market lows · refreshed every 60 s

pricing

Two ways to rent.
Pay only for the minutes you use.

Every server is priced by its host. These are the live floors across the marketplace — you'll see hundreds of variants once you're in.

Spot

$4.77 / hr
≈ 0.0000434 BTC · 3180 CLORE
  • Lowest possible rate
  • Per-minute billing
  • Can be interrupted by on-demand renter
  • Best for batch training, rendering
Browse spot B300s
MOST RENTED

On-demand

$5.80 / hr
≈ 0.0000527 BTC · 3867 CLORE
  • Guaranteed availability
  • No preemption, ever
  • Per-minute billing
  • Best for inference, dev work, demos
Rent on-demand
Pay with
Bitcoin on-chain
CLORE native token
USDT / USDC ERC-20 · BEP-20
workflow

Four steps to a running B300.

No sales call. No quota request. No three-week procurement. The first four commands are all you need.

01 / FILTER

Pick your card

Filter the marketplace by B300 288GB, country, GPU count, reliability score, network speed.

02 / RENT

Click rent

Choose a Docker image — PyTorch, vLLM, ComfyUI, Blender — or paste your own.

$ clore rent --gpu "B300 288GB"
03 / CONNECT

SSH or Jupyter

You get a public endpoint, an SSH key, and Jupyter on port 8888 in under 90 s.

04 / STOP

Stop anytime

Per-minute billing rounds to the second. Stop the instance and the meter stops with it.

faq

Questions hosts and renters ask.

How is B300 different from B200?

B300 is Blackwell Ultra — same FP4/FP8 tensor cores as B200 but with 288 GB HBM3e (50% more memory) and higher sustained throughput. The extra VRAM unlocks 1T+ parameter pretraining and 671B MoE serving with full KV cache headroom, where B200 had to swap experts. CLORE.AI lists B300 as supply ramps in 2026; available on bare-metal 8-GPU NVLink-Switch pods.

Can I pretrain a 70B model from scratch on this GPU?

Single-card, no - 70B pretraining needs an 8-GPU node minimum. CLORE.AI lists 8x H100, 8x H200, and 8x B200 pods with NVLink fabric for exactly this. A100 80GB pods run 70B FSDP training but at lower throughput than Hopper-class. For multi-week training, contact host operators for reserved-instance terms - listed in the marketplace under 'Reserved'.

What's the FP8 throughput here vs A100 80GB?

A100 80GB has no FP8 - peak is BF16/TF32. H100 introduces FP8 with TransformerEngine and roughly 4x the BF16 training throughput at 2x the rental price - so ~2x perf-per-dollar on FP8-eligible workloads. H200 matches H100 compute but adds 141 GB HBM3e. B200 doubles H100 FP8 again with 192 GB HBM3e. Pick by VRAM and bandwidth ceiling, not just sticker FLOPS.

Does this support NVLink-Switch / NVSwitch fabric?

8-GPU H100 SXM, H200 SXM, and B200 nodes ship with NVSwitch fabric - 900 GB/s peer bandwidth on H100/H200, 1.8 TB/s 5th-gen NVLink on B200. PCIe variants (H100 PCIe, A100 PCIe) have NVLink Bridge in pairs only. Multi-node fabric (NVLink-Switch across racks) is available on B200 hyperscale pods - filter by 'NVSwitch' in the marketplace.

Are 8-GPU pods available for FSDP and DeepSpeed training?

Yes. Multi-GPU listings expose all cards in a single rental as a coherent node with NVSwitch (where present), shared NVMe scratch, and InfiniBand or 100 GbE fabric for multi-node training. The standard PyTorch torchrun, DeepSpeed, and Megatron-LM launchers run unmodified. Filter the marketplace by GPU count to find 8x A100, 8x H100, 8x H200 nodes.

What's the HBM bandwidth comparison vs the predecessor?

V100 (HBM2, 900 GB/s) -> A100 40GB (HBM2e, 1,555 GB/s) -> A100 80GB (HBM2e, 1,935 GB/s) -> H100 (HBM3, 3,350 GB/s) -> H200 (HBM3e, 4,800 GB/s) -> B200 (HBM3e, 8,000 GB/s). Each generation roughly doubles bandwidth or VRAM; KV-cache-bound serving and bandwidth-bound training scale almost linearly with this number.

workload spotlight

Real numbers on the B300.

288 GB HBM3e + Blackwell Ultra FP4 — the trillion-parameter inference card. Pretrains 1T+ models and serves 671B MoE without expert swap.

Pretrain 1T+ MoE (8-GPU pod)
TransformerEngine FP4 + 5th-gen NVLink + FSDP
~16,000 tokens/s/GPU on 8× node

288 GB per card carries 1T+ parameter active weights without B200's expert-swap overhead. Frontier-lab unlock for trillion-scale pretraining.

Read the guide →
DeepSeek-V3 671B MoE serving
vLLM + expert parallelism + FP8/FP4
~5,200 tok/s aggregated on 8× node

288 GB per card serves 671B MoE with full KV cache headroom — no off-card expert spill, ~37% higher sustained throughput than B200.

Read the guide →
Llama-3 70B FP4 serving
vLLM + Blackwell Ultra FP4 + chunked prefill
~7,500 tok/s aggregated, 256 concurrent

Doubles B200 70B concurrency at the same per-request latency budget. The serving-throughput ceiling in 2026.

Read the guide →
datacenter comparison

Datacenter-tier comparison.

Side-by-side specs across the datacenter tier. Click any row to see that GPU.

GPU
HBM
Mem BW (GB/s)
FP8 TFLOPS
BF16 TFLOPS
NVLink BW
Transformer Engine
Spot $/hr
Tesla V100
32 GB HBM2
900
300 GB/s
$0.28
A100 40GB
40 GB HBM2e
1,555
312
600 GB/s
$0.78
A100 80GB
80 GB HBM2e
1,935
312
600 GB/s
$0.92
H100 tier focus
80 GB HBM3
3,350
1,979
989
900 GB/s
yes
$1.89
H200
141 GB HBM3e
4,800
1,979
989
900 GB/s
yes
$2.40
B200
192 GB HBM3e
8,000
~4,500
~2,250
1,800 GB/s
yes
$3.40
workload guides

Run these on your rented B300.

Step-by-step guides verified on CLORE.AI hardware. Pick a workload, copy the docker image, ship in minutes.

Training
DeepSpeed multi-GPU training
ZeRO-2/3 training across multiple cards.
Language Models
DeepSeek-V3
671B MoE model serving guide.
Language Models
vLLM serving
High-throughput LLM serving with PagedAttention.
Language Models
Llama 3.3 on CLORE.AI
Run Meta's flagship Llama-3.3 on your rented card.
Language Models
Qwen 2.5
Alibaba's Qwen 2.5 family.
Training
LLM fine-tuning
LoRA / QLoRA fine-tuning workflow.
Advanced
Multi-GPU setup
Configure NVLink, NCCL, and distributed training.
See all guides →
other gpus

Compare with similar cards.

B200
192 GB · from $3.4/hr
Rent →
H200
141 GB · from $1.55/hr
Rent →
H100
80 GB · from $1.49/hr
Rent →

Your training run
is 60 seconds away.

Texas and EU bare-metal hosts are accepting workloads right now. Sign up, top up your wallet, and the next hour is yours.