PRICING

Simple, transparent GPU pricing.

Pay per hour or commit for 6-12 months and save up to 30%. No egress fees on first 10TB. No hidden charges. Early access members lock in pricing for 12 months.

VENDOR

COMMITMENT

NVIDIA H100 80GB SXM

80 GB HBM3 · 989 TFLOPS (FP16)

Available

$1.8/GPU/hr

AWS$4

Azure$3.67

Lambda$2.49

CoreWeave$2.35

55% cheaper than AWS

NVIDIA B200

192 GB HBM3e · 2,250 TFLOPS (FP16)

Available

$2.5/GPU/hr

AWS$5.2

Azure$4.8

Lambda$3.8

CoreWeave$3.5

52% cheaper than AWS

AMD MI300X

192 GB HBM3 · 1,307 TFLOPS (FP16)

Available

$1.5/GPU/hr

AWS$3.5

Azure$3.2

Lambda$2.1

CoreWeave$2

57% cheaper than AWS

Huawei Ascend 910B

64 GB HBM2e · 640 TFLOPS (FP16)

Q1 2027

$1.2/GPU/hr

AWS—

Azure—

Lambda—

CoreWeave—

Cost Calculator

Accelerator

GPUs8x

Hours / day12h

Days / month20d

YOU SAVE PER YEAR

$50,688

NILE / MONTH

$3,456

AWS / MONTH

$7,680

Always Included

✓InfiniBand NDR 400Gbps interconnect
✓PyTorch, JAX, vLLM pre-configured
✓SSH, Jupyter, REST API access
✓Per-second billing

No Hidden Fees

✓No egress fees (first 10TB/mo)
✓No setup or teardown charges
✓No minimum commitment (on-demand)
✓No network transfer fees within cluster

Enterprise Add-ons

✓Dedicated account manager
✓Custom SLA with financial credits
✓Private cluster isolation
✓Priority capacity reservation

Join Early Access →

Early access pricing locked for 12 months. First 500 signups only.