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

Nebius

Amsterdam-headquartered, Nasdaq-listed AI cloud with NVIDIA H100 to B300 and GB200/GB300 capacity in Europe, the US, the UK and Israel, per-second billing, dynamic spot pricing and managed Kubernetes/Slurm; H100/H200/B200/B300 on-demand prices rise on 1 October 2026.

GPU types6with a current offer
Current offers6464 readings on file · last 45 days
H100 from$4.50per GPU-hour, on-demand
Entry price$1.35cheapest GPU-hour listed
Countries711 named regions
Trust Score v0.1
89/100
  • Company & track record73
  • Infrastructure control100
  • Pricing & billing transparency100
  • 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 disclosednebius.com · Sep 29, 20260 / 4
Headquarters publishedNebius B.V.nebius.com · Sep 29, 20263 / 3
Legal entity namedNebius B.V.nebius.com · Sep 29, 20264 / 4
Funding or ownership disclosedPublic, Nasdaq: NBISnebius.com · Sep 29, 20264 / 4
Infrastructure control
weight 15%
100 / 100
Capacity model publishedoperates or rents dedicated capacitynebius.com · Sep 29, 202610 / 10
Regions published11 regions or countriesnebius.com · Sep 29, 20265 / 5
Pricing & billing transparency
weight 20%
100 / 100
Public prices without an accountyes (read on the pricing page)nebius.com · Sep 29, 20265 / 5
Billing granularity publishedper secondnebius.com · Sep 29, 20264 / 4
Egress policy publishedincluded (VM egress/ingress free; Object Storage Standard egress 0.015 USD/GiB)nebius.com · Sep 29, 20264 / 4
Stopped-instance billing publishedCompute resources (GPU, vCPU, RAM) of stopped VMs are not charged; disks, snapshots and shared filesystems keep billingnebius.com · Sep 29, 20264 / 4
Public IPv4 policy publishedincludednebius.com · Sep 29, 20263 / 3
Reliability signals
weight 20%
75 / 100
Status pagehttps://status.nebius.com/nebius.com · Sep 29, 20266 / 6
SLA publishedCompute: 99.50% per VM; compensation 10% (99.00-99.50%), 15% (95-99%), 30% (below 95%) of the service feenebius.com · Sep 29, 20266 / 6
SLA with a numeric uptime commitmentCompute: 99.50% per VM; compensation 10% (99.00-99.50%), 15% (95-99%), 30% (below 95%) of the service feenebius.com · Sep 29, 20263 / 3
Uptime history publishednot disclosednebius.com · Sep 29, 20260 / 5
Observed availability (30 days)not measured yet—
Security & compliance
weight 15%
100 / 100
SOC 2Type IInebius.com · Sep 29, 20268 / 8
ISO 27001Listed on Trust Center (also ISO 22301, 27018, 27032, 27701, 27799, 42001)nebius.com · Sep 29, 20264 / 4
HIPAA or GDPR DPASOC 2 Type II with HIPAA; HIPAA Implementation Guideline publishednebius.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 →
Market report · Oct 25, 2026 · Specialised GPU clouds, hourly rental, pods and inference · #41 of the 200 to audit · prices read on the provider's page

What the market report notes about Nebius

Our internal report lists 248 actors with their zone, segment, status and the prices it could read, dated Oct 25, 2026. Its notes are reproduced as written, in French, with their sources.

Listed as
Nebius
Cross-checkfr
Hausse du 01/10/2026 : H100 4,50 $, H200 5,40 $, B200 8,50 $, B300 9,50 $ ; prix spot publics depuis le 22/10/2026 ; CA T2 2026 582 M$.
Prices by GPU · 6 GPU types with a current offer

What Nebius 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$4.50$3.17+42%2on-demandspotH100 prices
RTX PRO 6000$1.80$2.14−16%2on-demandspotRTX PRO 6000 prices
H200$5.40$4.76+13%2on-demandspotH200 prices
L40S$1.35$1.26+7%4on-demandspotL40S prices
B200$8.50$7.50+13%2on-demandspotB200 prices
B300$9.50$8.96+6%2on-demandspotB300 prices
All offers · 64 readings on file

Every Nebius 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
L40S1× L40SSpotNebius regions (FinlandStock unknown$0.65$0.6521:09 UTCfreshReceipt
L40S1× L40SSpotNebius regions (FinlandStock unknown$0.65$0.6521:09 UTCfreshReceipt
H200SXMoutlier · suspended1× H200 SXMSpotNebius regions (FinlandStock unknown$0.79$0.7921:09 UTCfreshReceipt
H100SXM1× H100 SXMSpotNebius regions (FinlandStock unknown$0.79$0.7921:09 UTCfreshReceipt
RTX PRO 60001× RTX PRO 6000SpotNebius regions (FinlandStock unknown$0.79$0.7921:09 UTCfreshReceipt
B300SXMoutlier · suspended1× B300 SXMSpotNebius regions (FinlandStock unknown$0.99$0.9921:09 UTCfreshReceipt
B200SXMoutlier · suspended1× B200 SXMSpotNebius regions (FinlandStock unknown$0.99$0.9921:09 UTCfreshReceipt
L40S1× L40SOn-demandNebius regions (FinlandStock unknown$1.35$1.3521:09 UTCfreshReceipt
L40S1× L40SOn-demandNebius regions (FinlandStock unknown$1.35$1.3521:09 UTCfreshReceipt
RTX PRO 60001× RTX PRO 6000On-demandNebius regions (FinlandStock unknown$1.80$1.8021:09 UTCfreshReceipt
H100SXM1× H100 SXMOn-demandNebius regions (FinlandStock unknown$4.50$4.5021:09 UTCfreshReceipt
H200SXM1× H200 SXMOn-demandNebius regions (FinlandStock unknown$5.40$5.4021:09 UTCfreshReceipt
B200SXM1× B200 SXMOn-demandNebius regions (FinlandStock unknown$8.50$8.5021:09 UTCfreshReceipt
B300SXM1× B300 SXMOn-demandNebius regions (FinlandStock unknown$9.50$9.5021:09 UTCfreshReceipt
14 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 second
Minimum charge
not published
Stopped instance
Compute resources (GPU, vCPU, RAM) of stopped VMs are not charged; disks, snapshots and shared filesystems keep billing
Egress
included (VM egress/ingress free; Object Storage Standard egress 0.015 USD/GiB)
Public IPv4
included
Billing model
both
Minimum deposit
25 USD charged when the card is added and credited to the balance; manual top-ups from 25 USD
Free credits
AI Builder Program advertised on the pricing page: '$400+ in credits and discounts'; startup program and research grants

Account

Payment
card · invoice · crypto (not published)
Sign-up
Google, GitHub or Microsoft account, name, email, card, billing address, company tax/VAT ID (companies)
Identity check
card verification
Support
tickets, email
Enterprise support
not published
SLA
Compute: 99.50% per VM; compensation 10% (99.00-99.50%), 15% (95-99%), 30% (below 95%) of the service fee
Published uptime
not published

Infrastructure

Regions
11 named regions in 7 countrieseu-north1 (Finland) · eu-south1 (Spain) · eu-west1 (France) · eu-west2 (France) · me-west1 (Israel) · us-central1 (United States) · us-north1 (United States) · uk-south1 (United Kingdom) · uk-south2 (United Kingdom) · eu-north2 (private · existing deployments only) (Iceland)
Countries
Finland, Spain, France, Israel, United States, United Kingdom, Iceland
Platform
api, cli, serverless, ssh, custom_docker_image, persistent_volumes, snapshots, spot_or_interruptible · SDK : go, python, typescript (gRPC API) · Terraform : official · Kubernetes : managed · interconnexion : InfiniBand 400 Gbps per GPU (8x ConnectX-7) on H100/H200/B200, 800 Gbps per GPU (ConnectX-8) on B300 · NVL72 for GB200/GB300 · clusters : GPU clusters of 8-GPU VMs over InfiniBand · Managed Soperator (Slurm on Kubernetes) free · GB200/GB300 NVL72 on request
Storage
volume · 0.071 USD/GiB-month (Network SSD, 730 h) · Block volumes with erasure coding on pricing pagevolume · 0.053 USD/GiB-month (Network SSD non-replicated) · Nonevolume · 0.118 USD/GiB-month (Network SSD IO M3, 3x mirroring) · Nonevolume · 0.065 USD/GiB-month (local SSD) · Write endurance limit: no more than disk capacity per daynetwork · 0.08 USD/GiB-month (Shared Filesystem SSD) · WEKA filesystem 0.10 USD/GiB-month on pricing pageobject · 0.0147 USD/GiB-month (Standard) + 0.015 USD/GiB egress · Intelligent warm 0.030, cold 0.010 (egress free); Enhanced 0.11 (egress free)
Migration out
VMs use standard images and Docker containers; data can be exported from disks, shared filesystems and S3-compatible Object Storage. VM egress is free, but Standard-tier Object Storage egress is billed at 0.015 USD/GiB. Terraform and Kubernetes manifests port to other clouds.

Security & compliance

SOC 2
Type II
ISO 27001
Listed on Trust Center (also ISO 22301, 27018, 27032, 27701, 27799, 42001)
HIPAA
SOC 2 Type II with HIPAA; HIPAA Implementation Guideline published
GDPR / DPA
Data Processing Agreement published; GDPR and CCPA policies
Other
SOC 3, DORA, CSA STAR Level 1
Legal entity
Nebius B.V.
Funding
Public, Nasdaq: NBIS
GPUBank market study, read on the provider's own pages · 19 pages read · facts marked "not published" were looked for and not foundPages read ↓
From the study · assessment

Nebius, in short

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

Strengths

  • Newest NVIDIA hardware self-serve: B300 and B200 per GPU-hour, GB200/GB300 NVL72 on request
  • Per-second billing; stopped VMs pay storage only
  • Egress, ingress and public IPs free
  • Dynamic spot pricing with advertised floors from 0.79-0.99 USD per GPU-hour and user-set max-price policies
  • Managed Kubernetes and Managed Soperator (Slurm) free of charge
  • Nine public regions across Finland, France, Spain, UK, Israel and two US sites
  • Broad compliance set: SOC 2 Type II with HIPAA, SOC 3, ISO 27001/27701/42001 and more
  • Official CLI, Terraform provider and Go/Python/TypeScript SDKs

Weaknesses

  • On-demand H100 at 3.85 USD per GPU-hour rising to 4.50 on 1 October 2026; H200, B200 and B300 rise by 17-21%
  • Each GPU model is limited to a few regions (H100 only in Finland)
  • Preemptible price information conflicts between the pricing page (dynamic, from 0.79) and docs (fixed 2.15 for H100)
  • Commitment discounts (up to 35%) only for company accounts through sales
  • Bank transfer only for companies; no crypto published
  • Compute SLA 99.5% with compensation capped at 30%

Ideal for

  • Teams needing B200/B300 or large InfiniBand clusters with managed Kubernetes or Slurm
  • European buyers requiring EU regions and a wide compliance portfolio
  • Interruptible batch jobs that can exploit dynamic spot pricing

Not published

  • Current spot prices per platform (shown only in the console)
  • Which preemptible figure applies at checkout: dynamic spot from 0.79 (pricing page) or fixed 2.15 for H100 (docs)
  • Commitment discount rates and terms
  • GB200 and GB300 NVL72 prices (contact sales)
  • Crypto payment (not mentioned)
  • Whether identity documents are requested beyond card verification
  • Bare-metal availability
  • Founding year (not on pages read)

Observations

Pricing page shows two on-demand columns: current and 'Effective October 1, 2026'

Spot prices are shown as 'from' floors with a footnote that they change as often as every 15 minutes

Promotions for the AI Builder Program ($400+ credits) and Physical AI Awards ($150K compute) on the banner

Payment FAQ says cards through Stripe; bank transfers also available

Pages read

Questions about Nebius

What GPUs does Nebius offer?

Nebius has 6 GPU types with a fresh offer today, for example H100 from $4.50, H200 from $5.40, RTX PRO 6000 from $1.80, L40S from $1.35, B200 from $8.50, B300 from $9.50 per GPU-hour on-demand. The full list with prices and market medians is on this page.

How does Nebius bill?

Billing granularity: per second. Stopped instance: Compute resources (GPU, vCPU, RAM) of stopped VMs are not charged; disks, snapshots and shared filesystems keep billing

How trustworthy is Nebius?

Its Trust Score is 89/100 (v0.1), read from its own documents. Strong on infrastructure control, pricing & billing transparency, security & compliance. The score measures transparency and published guarantees, not customer satisfaction; every criterion and its source is listed above.

Where are Nebius's data centres?

11 named regions in 7 countries: Finland, Spain, France, Israel, United States, United Kingdom, Iceland.

How do you pay Nebius, and are there credits?

Payment by card, invoice, crypto. Free credits: AI Builder Program advertised on the pricing page: '$400+ in credits and discounts'; startup program and research grants. Identity check: card verification.