gpurank
esc
  • Nothing named “”. Press Enter to search the offers instead.
↑↓ to move · ↵ to open · esc to close0 entries
Operator · Livingston, NJ

CoreWeave

US AI cloud selling Kubernetes-native bare-metal NVIDIA GPU nodes (mostly 8-GPU HGX nodes and GB200/GB300 NVL72) on-demand, spot and reserved, aimed at AI labs and enterprises.

GPU types18with a current offer
Current offers105105 readings on file · last 45 days
H100 from$6.16per GPU-hour, on-demand
Entry price$1.25cheapest GPU-hour listed
Countries239 named regions
Trust Score v0.1
85/100
  • Company & track record73
  • Infrastructure control100
  • Pricing & billing transparency80
  • Reliability signals75
  • Security & compliance100
  • Customer signal—

Scored Sep 29, 2026 from the provider's own pages. Dimensions not yet measured are excluded from the denominator. Criteria ↓

Trust Score · 5 of 6 dimensions measured

The score, criterion by criterion

Each point comes from something the provider publishes, with the page it was read on and the date. Anything not disclosed earns zero.

Company & track record
weight 15%
73 / 100
Founding year publishednot disclosedcoreweave.com · Sep 29, 20260 / 4
Headquarters publishedCoreWeave, Inc.coreweave.com · Sep 29, 20263 / 3
Legal entity namedCoreWeave, Inc.coreweave.com · Sep 29, 20264 / 4
Funding or ownership disclosedInvestor Relations site linked from the footer (investors.coreweave.com returned a bot challenge); listing details not read on a primary page todaycoreweave.com · Sep 29, 20264 / 4
Infrastructure control
weight 15%
100 / 100
Capacity model publishedoperates or rents dedicated capacitycoreweave.com · Sep 29, 202610 / 10
Regions published39 regions or countriescoreweave.com · Sep 29, 20265 / 5
Pricing & billing transparency
weight 20%
80 / 100
Public prices without an accountyes (read on the pricing page)coreweave.com · Sep 29, 20265 / 5
Billing granularity publishedper hourcoreweave.com · Sep 29, 20264 / 4
Egress policy publishedincluded (data transfer between CoreWeave and internet: Free; storage egress and IOPS: Free)coreweave.com · Sep 29, 20264 / 4
Stopped-instance billing publishednot disclosedcoreweave.com · Sep 29, 20260 / 4
Public IPv4 policy publishedextra: 4.00 USD per public IP per month; Bring Your Own IP 4.00 USD one-time setup per IP; NAT Gateway freecoreweave.com · Sep 29, 20263 / 3
Reliability signals
weight 20%
75 / 100
Status pagehttps://status.coreweave.com/coreweave.com · Sep 29, 20266 / 6
SLA publishedSLO in Terms of Service: >=99.9% monthly uptime for instances in multiple regions, >=99% for a single region; credits 10-50% of the monthly bill, capped at 50%; not applicable to Spotcoreweave.com · Sep 29, 20266 / 6
SLA with a numeric uptime commitmentSLO in Terms of Service: >=99.9% monthly uptime for instances in multiple regions, >=99% for a single region; credits 10-50% of the monthly bill, capped at 50%; not applicable to Spotcoreweave.com · Sep 29, 20263 / 3
Uptime history publishednot disclosedcoreweave.com · Sep 29, 20260 / 5
Observed availability (30 days)not measured yet—
Security & compliance
weight 15%
100 / 100
SOC 2Type IIcoreweave.com · Sep 29, 20268 / 8
ISO 27001yes (Trust Center badge)coreweave.com · Sep 29, 20264 / 4
HIPAA or GDPR DPAData Processing Agreement referenced in Terms of Servicecoreweave.com · Sep 29, 20263 / 3
Customer signal
weight 15%
not measured
Aggregated public reviewsnot collected yet—
provisional v0.1: evidence from the provider's own pages; dimensions not yet measured are excluded from the denominatorFull method →
Prices by GPU · 9 GPU types with a current offer

What CoreWeave charges, against the market

Cheapest current on-demand true cost here, next to the market median for the same GPU. Negative means cheaper than the market.

GPUCheapest hereMarket medianvs marketCurrent offersModes
H100$6.16$3.17+94%3on-demandspotH100 prices
A100 80GB$2.70$1.92+41%3on-demandspotA100 80GB prices
H200$6.31$4.80+31%3on-demandspotH200 prices
L40S$2.25$1.24+82%2on-demandspotL40S prices
RTX PRO 6000$2.50$2.15+17%3on-demandspotRTX PRO 6000 prices
B200$8.60$7.50+15%3on-demandspotB200 prices
L40$1.25$0.88+41%2on-demandspotL40 prices
GH200$6.50$2.75+137%1on-demandGH200 prices
GB200$10.50$8.00+31%1on-demandGB200 prices
All offers · 105 readings on file

Every CoreWeave offer, with its receipt

Current offers (last 45 days) ranked by true cost, all rental modes, each with its read date. Older readings stay listed below, in grey.

GPUConfigurationModeRegionStockListTrue costRead
L408× L40Spot—Stock unknown$0.78$0.7800:32 UTCfreshReceipt
L40S8× L40SSpot—Stock unknown$0.99$0.9900:32 UTCfreshReceipt
A100 80GB8× A100 80GBSpot—Stock unknown$1.19$1.1900:32 UTCfreshReceipt
A100 80GB8× A100 80GBSpot—Stock unknown$1.21$1.2100:32 UTCfreshReceipt
L408× L40On-demand—Stock unknown$1.25$1.2500:32 UTCfreshReceipt
RTX PRO 60008× RTX PRO 6000Spot—Stock unknown$1.38$1.3800:32 UTCfreshReceipt
RTX PRO 60008× RTX PRO 6000Spot—Stock unknown$1.39$1.3900:32 UTCfreshReceipt
L40S8× L40SOn-demand—Stock unknown$2.25$2.2500:32 UTCfreshReceipt
H100SXM8× H100 SXMSpot—Stock unknown$2.44$2.4400:32 UTCfreshReceipt
H100SXM8× H100 SXMSpot—Stock unknown$2.46$2.4600:32 UTCfreshReceipt
RTX PRO 60008× RTX PRO 6000On-demand—Stock unknown$2.50$2.5000:32 UTCfreshReceipt
H2008× H200Spot—Stock unknown$2.58$2.5800:32 UTCfreshReceipt
H2008× H200Spot—Stock unknown$2.62$2.6200:32 UTCfreshReceipt
A100 80GB8× A100 80GBOn-demand—Stock unknown$2.70$2.7000:32 UTCfreshReceipt
B2008× B200Spot—Stock unknown$4.26$4.2600:32 UTCfreshReceipt
B2008× B200Spot—Stock unknown$4.36$4.3600:32 UTCfreshReceipt
H100SXM8× H100 SXMOn-demand—Stock unknown$6.16$6.1600:32 UTCfreshReceipt
H2008× H200On-demand—Stock unknown$6.31$6.3100:32 UTCfreshReceipt
GH2001× GH200On-demand—Stock unknown$6.50$6.5000:32 UTCfreshReceipt
B2008× B200On-demand—Stock unknown$8.60$8.6000:32 UTCfreshReceipt
GB2004× GB200On-demand—Stock unknown$10.50$10.5000:32 UTCfreshReceipt
21 fresh offers · prices per GPU-hour · true cost includes the add-ons the provider publishesOpen in the comparison tool →
True cost & billing · read Sep 29, 2026

What the hourly price does not say

Billing granularity, minimums, what a stopped machine costs, traffic, addresses, deposits and identity checks, as published by the provider.

Billing

Billing granularity
per hour
Minimum charge
not published
Stopped instance
not published
Egress
included (data transfer between CoreWeave and internet: Free; storage egress and IOPS: Free)
Public IPv4
extra: 4.00 USD per public IP per month; Bring Your Own IP 4.00 USD one-time setup per IP; NAT Gateway free
Billing model
postpaid invoice
Minimum deposit
not published
Free credits
not published

Account

Payment
card · invoice · crypto (not published)
Sign-up
business email, company, sales contact form
Identity check
sales contract
Support
email ([email protected]), support tickets, sales contact form
Enterprise support
not published
SLA
SLO in Terms of Service: >=99.9% monthly uptime for instances in multiple regions, >=99% for a single region; credits 10-50% of the monthly bill, capped at 50%; not applicable to Spot
Published uptime
not published

Infrastructure

Regions
39 named regions in 2 countriesUS-EAST-01 (United States) · US-EAST-02 (United States) · US-EAST-03 (United States) · US-EAST-04 (United States) · US-EAST-06 (United States) · US-EAST-08 (United States) · US-EAST-11 (United States) · US-EAST-12 (United States) · US-EAST-13 (United States) · US-EAST-14 (United States) · US-EAST-15 (United States) · US-EAST-18 (United States) · US-CENTRAL-01 (United States) · US-CENTRAL-02 (United States) · US-CENTRAL-05 (United States) · US-CENTRAL-06 (United States) · US-CENTRAL-07 (United States) · US-CENTRAL-08 (United States) · US-CENTRAL-09 (United States) · US-CENTRAL-10 (United States) · US-CENTRAL-11 (United States) · US-WEST-01 (United States) · US-WEST-02 (United States) · US-WEST-03 (United States) · US-WEST-04 (United States) · US-WEST-07 (United States) · US-WEST-09 (United States) · US-WEST-10 (United States) · US-ORD1 (United States) · US-LAS1 (United States) · US-LGA1 (United States) · US-RNO2 (United States) · RDU1 (United States) · CA-EAST-01 (Canada) · EU-SOUTH-03 (None) · EU-SOUTH-04 (None) · EU-NORTH-02 (None) · EU-NORTH-04 (None) · EU-NORTH-05 (None)
Countries
United States, Canada
Platform
API, serverless, custom Docker images, persistent volumes, spot or interruptible, bare metalTerraform: officialKubernetes: managedInterconnect: NVIDIA Quantum-2 InfiniBand 3200 Gbps (H200 section of GPU Compute page); NVLink inside GB200/GB300 NVL72 racksClusters: CKS managed Kubernetes on bare metal (control plane free), SUNK Slurm on Kubernetes (free excluding CPU usage), GB200/GB300 NVL72 racks, Direct Connect 10G-400G
Storage
object · 0.06 USD/GB-month Hot, 0.03 Warm, 0.015 Cold, 0.0125 Archive (limited availability) · CoreWeave AI Object Storage; GB = GiBnetwork · 0.070 USD/GB-month · Distributed File Storagenetwork · File/Block HDD 0.04, NVMe 0.07 USD/GB-month; Object 0.03 USD/GB-month · Classic pricing page (legacy)
Migration out
Workloads run as containers on CKS (standard Kubernetes), so manifests and images move to any Kubernetes host. AI Object Storage is S3-compatible and egress is free, so data export has no transfer fee. Terms of Service ask customers to export data before account closure; deletion follows a recovery period of up to 30 days.

Security & compliance

SOC 2
Type II
ISO 27001
yes (Trust Center badge)
HIPAA
not published
GDPR / DPA
Data Processing Agreement referenced in Terms of Service
Other
ISO 27017 and ISO 27018 badges on Trust Center; EU Data Act Addendum listed in docs policies
Legal entity
CoreWeave, Inc.
Funding
Investor Relations site linked from the footer (investors.coreweave.com returned a bot challenge); listing details not read on a primary page today
Read on the provider's own pages · 15 pages read · facts marked "not published" were looked for and not foundPages read ↓
Assessment · documents read Sep 29, 2026

CoreWeave, in short

What stood out when we read the documents: strengths, weaknesses, who it suits, and what we could not find.

Strengths

  • Broadest published NVIDIA Blackwell range of this set: GB200 NVL72 42.00 USD per instance-hour (10.50 per GPU), HGX B200 68.80 USD per node-hour (8.60 per GPU), HGX B300 spot from 35.84 USD per node-hour.
  • Spot prices well under on-demand, e.g. HGX H100 spot 19.71 USD per node-hour in North America (2.46 per GPU) vs 49.24 on-demand.
  • No charge for internet egress, intra-CoreWeave traffic, NAT gateway, VPC, CKS control plane or SUNK.
  • Kubernetes-native bare metal with an official Terraform provider (v0.24.0, published 2026-09-11) and public OpenAPI specs.
  • Trust Center shows SOC 2 Type II, ISO 27001, ISO 27017 and ISO 27018 badges.
  • Written SLO with service credits (99.9% multi-region, 99% single region).

Weaknesses

  • On-demand GPUs sold as 8-GPU nodes; no single-GPU on-demand instance except GH200 (per-GPU inference prices only for inference platform customers).
  • On-demand HGX H100 normalises to 6.155 USD per GPU-hour, HGX H200 to 6.305 and HGX B200 to 8.60.
  • GB300 NVL72, HGX B300 on-demand and RTX PRO 6000 Standard Memory on-demand are 'Contact sales'.
  • Public IPv4 billed at 4.00 USD per IP per month.
  • Docs for regions, instances and billing require an access code; onboarding goes through a sales form asking for business email and company.
  • Postpaid monthly invoicing; invoices of 20,000 USD or more must be paid by ACH or wire, Net 15; no crypto option published.

Ideal for

  • AI labs and enterprises that need multi-node HGX or NVL72 clusters with InfiniBand, managed Kubernetes or Slurm, and a contract
  • Teams able to run fault-tolerant jobs on 8-GPU spot nodes

Not published

  • Billing granularity (per hour vs per second or minute) not stated on public pages
  • the billing docs need an access code. 'per hour' reflects the published price unit only.
  • Stopped-instance billing, minimum charge and free credits not published publicly.
  • Region-to-city and EU region countries: 'Regions and Availability Zones' docs page needs an access code
  • Reserved pricing: only 'up to 60% discounts' published, no per-term prices.
  • A100 form factor (SXM vs PCIe) on the current pricing page not stated
  • Whether classic a la carte pricing is still orderable by new customers not stated.
  • CLI, SDK, Jupyter and SSH access not verified on public pages.
  • Founding year and listing or funding status not read on an accessible primary page (investor site blocked).
  • Identity verification beyond the sales process not published
  • HIPAA status not published.
  • Payment methods: Terms mention credit card and ACH/wire
  • PayPal and crypto not mentioned.

Observations

Pricing page shows hourly per-instance prices for North America and Europe without an account

Prices are per 8-GPU node, except an 'Inference Single GPU Price' column restricted to inference platform customers

Reserved capacity only described as 'up to 60% discounts' with a contact request

A separate classic pricing page still lists a la carte per-GPU component prices for older GPUs (A40, RTX A6000, V100, Quadro RTX)

Docs pages for regions, instances and billing ask for an access code

Pages read

Questions about CoreWeave

What GPUs does CoreWeave offer?

CoreWeave has 18 GPU types with a fresh offer today, for example H100 from $6.16, A100 80GB from $2.70, H200 from $6.31, RTX PRO 6000 from $2.50, L40S from $2.25, B200 from $8.60 per GPU-hour on-demand. The full list with prices and market medians is on this page.

How does CoreWeave bill?

Billing granularity: per hour.

How trustworthy is CoreWeave?

Its Trust Score is 85/100 (v0.1), read from its own documents. Strong on infrastructure control, security & compliance. The score measures transparency and published guarantees, not customer satisfaction; every criterion and its source is listed above.

Where are CoreWeave's data centres?

39 named regions in 2 countries: United States, Canada.

How do you pay CoreWeave, and are there credits?

Payment by card, invoice, crypto. No sign-up credit is published. Identity check: sales contract.