CoreWeave

CuratedActive

coreweave.com

United States, LivingstonFounded 2017

Largest independent AI neocloud, renting NVIDIA GB300, B200 and H200 clusters

GPU cloud

Visit website Pricing

From

—

GPU from

$2.25/GPU-hr

Locations

countries

Data centers

sites

Networks (ASN)

ASNs

Plans & prices

plans

Overview

CoreWeave runs a Kubernetes-based cloud built for GPU workloads, renting NVIDIA HGX and GB200/GB300 NVL72 systems with InfiniBand fabrics to AI labs and enterprises. Founded in 2017 as a crypto-mining operation, it listed on Nasdaq (CRWV) in 2025 and bought Weights & Biases the same year.

Plans & prices

GPU instances8

Instance	Config	per GPU-hr	Total /hr	Spot	Source
NVIDIA L40S — 8-GPU instance at $18.00/h, normalized per GPU	1× L40S	$2.25	$2.25	—	Verified 3 days ago
RTX PRO 6000 Blackwell Server Edition (high memory) — 8-GPU instance at $20.00/h, normalized per GPU	1× RTX PRO 6000 Blackwell	$2.50	$2.50	—	Verified 3 days ago
NVIDIA A100 80GB — 8-GPU instance at $21.60/h, normalized per GPU	1× A100 80GB	$2.70	$2.70	—	Verified 3 days ago
NVIDIA HGX H100 — 8-GPU instance at $49.24/h, normalized per GPU	1× H100 SXM	$6.16	$6.16	—	Verified 3 days ago
NVIDIA HGX H200 — 8-GPU instance at $50.44/h, normalized per GPU	1× H200	$6.31	$6.31	—	Verified 3 days ago
NVIDIA GH200 — 1-GPU instance	1× GH200 (Grace Hopper)	$6.50	$6.50	—	Verified 3 days ago
NVIDIA HGX B200 — 8-GPU instance at $68.80/h, normalized per GPU	1× B200	$8.60	$8.60	—	Verified 3 days ago
NVIDIA GB200 NVL72 — 4-GPU instance at $42.00/h, normalized per GPU	1× GB200 (Grace Blackwell)	$10.50	$10.50	—	Verified 3 days ago