CUDO Compute
UK-run GPU cloud that now markets quote-only enterprise clusters (H100 to GB300 and Vera Rubin) while keeping a self-serve console and API with live per-GPU-hour prices in two Nordic data centers.
Trust Score v0.1- Company & track record47
- 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 measuredThe 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.
| Founding year published | not disclosedcudocompute.com · Sep 29, 2026 | 0 / 4 |
| Headquarters published | Cudo Ventures Ltdcudocompute.com · Sep 29, 2026 | 3 / 3 |
| Legal entity named | Cudo Ventures Ltdcudocompute.com · Sep 29, 2026 | 4 / 4 |
| Funding or ownership disclosed | not disclosedcudocompute.com · Sep 29, 2026 | 0 / 4 |
| Capacity model published | operates or rents dedicated capacitycudocompute.com · Sep 29, 2026 | 10 / 10 |
| Regions published | 2 regions or countriescudocompute.com · Sep 29, 2026 | 5 / 5 |
| Public prices without an account | yes (read on the pricing page)cudocompute.com · Sep 29, 2026 | 5 / 5 |
| Billing granularity published | per secondcudocompute.com · Sep 29, 2026 | 4 / 4 |
| Egress policy published | not disclosedcudocompute.com · Sep 29, 2026 | 0 / 4 |
| Stopped-instance billing published | All billable resources keep accruing charges while a VM is stopped or powered off; resources must be deleted to stop billingcudocompute.com · Sep 29, 2026 | 4 / 4 |
| Public IPv4 policy published | extra: 0.0035 USD/hour in no-kristiansand-1; 0 in se-skarholmen-1. Virtual networks 0.00454 USD/hourcudocompute.com · Sep 29, 2026 | 3 / 3 |
| Status page | https://status.cudocompute.com/cudocompute.com · Sep 29, 2026 | 6 / 6 |
| SLA published | Committed Resources SLA: 99.5% monthly uptime for subscription services with a committed period; service credits as a percentage of fees; no SLA published for on-demandcudocompute.com · Sep 29, 2026 | 6 / 6 |
| SLA with a numeric uptime commitment | Committed Resources SLA: 99.5% monthly uptime for subscription services with a committed period; service credits as a percentage of fees; no SLA published for on-demandcudocompute.com · Sep 29, 2026 | 3 / 3 |
| Uptime history published | not disclosedcudocompute.com · Sep 29, 2026 | 0 / 5 |
| Observed availability (30 days) | not measured yet | — |
| SOC 2 | Type II listed on the home page; About page says 'SOC 2-aligned controls'; no report link foundcudocompute.com · Sep 29, 2026 | 8 / 8 |
| ISO 27001 | ISO 27001-certified facilities (home and About pages); About page also lists ISO 14001cudocompute.com · Sep 29, 2026 | 4 / 4 |
| HIPAA or GDPR DPA | Data Processing Addendum listed as a contract document 'where applicable'cudocompute.com · Sep 29, 2026 | 3 / 3 |
| Aggregated public reviews | not collected yet | — |
What the market report notes about CUDO Compute
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
- CUDO Compute
- Cross-checkfr
- H100 1,79 $ (agrégateur).
What CUDO Compute 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.
| GPU | Cheapest here | Market median | vs market | Current offers | Modes | |
|---|---|---|---|---|---|---|
| H100 | $1.82 | $3.17 | −43% | 1 | on-demand | H100 prices |
| L40S | $0.88 | $1.26 | −30% | 1 | on-demand | L40S prices |
Every CUDO Compute 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.
| GPU | Configuration | Mode | Region | Stock | List | True cost | Read | |
|---|---|---|---|---|---|---|---|---|
| L40S | 1× L40S1 vCPU · 1 GB RAM | On-demand | Norway | In stock | $0.88 | $0.88 | 21:09 UTCfresh | Receipt |
| H100SXM | 1× H100 SXM1 vCPU · 4 GB RAM | On-demand | Norway | Unavailable | $1.82 | $1.82 | 21:09 UTCfresh | Receipt |
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
- All billable resources keep accruing charges while a VM is stopped or powered off; resources must be deleted to stop billing
- Egress
- not published
- Public IPv4
- extra: 0.0035 USD/hour in no-kristiansand-1; 0 in se-skarholmen-1. Virtual networks 0.00454 USD/hour
- Billing model
- prepaid credits
- Minimum deposit
- 10 USD minimum credit to deploy a billable resource; a single top-up must be between 10 and 1000 USD
- Free credits
- not published
Account
- Payment
- card · PayPal · crypto (not published)
- Sign-up
- email (verified) and password, or GitHub/Google/Microsoft login, billing account: full name, country, address (organizations: legal name, tax ID), card or PayPal to add credit
- Identity check
- not published
- Support
- support ticket in console, contact form (sales)
- Enterprise support
- not published
- SLA
- Committed Resources SLA: 99.5% monthly uptime for subscription services with a committed period; service credits as a percentage of fees; no SLA published for on-demand
- Status page
- status.cudocompute.com ↗
- Published uptime
- not published
Infrastructure
- Regions
2 named regions in 2 countries
no-kristiansand-1 (Norway) · se-skarholmen-1 (Sweden)- Countries
- Norway, Sweden
- Platform
- api, cli, jupyter, ssh, persistent_volumes, snapshots, bare_metal · SDK : python · Terraform : official · Kubernetes : no · interconnexion : InfiniBand on cluster machine types · enterprise clusters with NVIDIA Quantum-X800 InfiniBand or Spectrum-X Ethernet · clusters : Self-serve clusters of 8-GPU bare-metal machines via API · enterprise clusters from 128 to 40,000 GPUs (clusters page) or 8 to 1,000+ GPUs (product pages), quote only
- Storage
- volume · 0.00014 USD/GiB-hour (0.1008 USD/GiB-month at 720 h) · VM disk pool price in both data centers (API diskGibPriceHr)object · 0.00007 USD/GiB-hour (0.0504 USD/GiB-month at 720 h) in no-kristiansand-1; 0.00013 USD/GiB-hour (0.0936 USD/GiB-month at 720 h) in se-skarholmen-1 · API objectStorageGibPriceHrnetwork · not published · NFS volumes exist; prices are behind the authenticated volumes/prices endpoint
- Migration out
- VM boot disks can be saved as private images and data disks detached; object storage is S3-compatible (s3cmd/s5cmd documented), so data can be copied out with standard S3 tools. Deleting a VM loses its boot disk; machines within a commitment term cannot be deleted.
Security & compliance
- SOC 2
- Type II listed on the home page; About page says 'SOC 2-aligned controls'; no report link found
- ISO 27001
- ISO 27001-certified facilities (home and About pages); About page also lists ISO 14001
- HIPAA
- not published
- GDPR / DPA
- Data Processing Addendum listed as a contract document 'where applicable'
- Other
- Terms governed by the laws of England and Wales
- Legal entity
- Cudo Ventures Ltd
- Funding
- not published
CUDO Compute, in short
What stood out when we read the documents: strengths, weaknesses, who it suits, and what we could not find.
Strengths
- H100 SXM GPU component at 1.79 USD/GPU-hour on demand (VM) and 8x H100 SXM or 8x H100 NVL bare metal at 14.32 USD/hour (1.79 per GPU)
- Per-second usage metering on prepaid credit
- Public, keyless API returns live GPU, vCPU, RAM, disk and IPv4 prices plus free-capacity counts
- Official Terraform provider (CudoVentures/cudo), cudoctl CLI and Python client
- Card and PayPal accepted; minimum top-up 10 USD
- Enterprise catalogue includes B300, GB300 NVL72 and Vera Rubin NVL72 under contract
Weaknesses
- Marketing pricing page shows 'Quote on request' for every GPU; self-serve prices are only visible through the console or API
- GPU prices exclude vCPU, RAM, disk and IPv4, which are metered separately
- Stopped VMs keep billing for all resources; resources are deleted when the credit balance reaches 0
- Only two self-serve data centers (Norway, Sweden) returned by the public API, with small free capacity (4 H100 VM GPUs, 12 A100, 3 L40S at capture time)
- SLA covers committed resources only (99.5%)
- No spot, serverless or managed Kubernetes product
Ideal for
- Teams wanting cheap H100 or L40S capacity in the Nordics with API or Terraform automation
- Enterprises negotiating dedicated Blackwell or Rubin clusters with managed operations
Not published
- Commitment (reserved) prices: the API supports purchasing VM commitments but the public machine-types endpoints only return COMMITMENT_TERM_NONE prices
- Enterprise cluster prices for all 8 pricing-page GPUs and B100: quote only
- 13 GPU models listed by the public gpu-models API (MI300X, A40, A800 PCIe, H200 NVL, H200 SXM, RTX 3080, RTX 4000 SFF Ada, RTX A4000, RTX A4000 Ada, RTX A4500, RTX A5000, RTX A6000, V100) had no machine type or price on 2026-09-29
- they are counted in catalogue_count
- NFS volume prices: authenticated endpoint only
- Egress price: not published
- Crypto payment: not mentioned in the payment methods page
- Identity verification beyond email and billing details: not documented
- HQ city and country, founding year, funding: not stated on the pages read (terms under English law)
- Full list of the '16 ISO-certified data centres' used for enterprise clusters: not published
- SOC 2: home page lists 'SOC 2 Type II' while the About page says 'SOC 2-aligned'
- no report seen
Observations
Pricing page title is 'Commercial models for enterprise AI infrastructure' and lists 8 GPUs, all 'Quote on request'
Each GPU product page shows 'Starting from: Contact us for pricing'
No link to the self-serve console or its prices from the marketing site; the docs still describe the on-demand platform
Footer shows a status indicator ('All Systems Operational')
Pages read
- cudocompute.com/pricing — Enterprise GPU list, quote on request
- cudocompute.com/ — Positioning, regions, SOC 2 Type II and ISO 27001 claims, capacity pipeline
- cudocompute.com/about-us — Leadership, compliance wording, operating regions
- cudocompute.com/products/nvidia-h100 — H100 node specs, 16 ISO-certified data centres, 8 to 1,000+ GPUs
- cudocompute.com/products/nvidia-h200 — H200 specs
- cudocompute.com/products/nvidia-b200 — B200 specs
- cudocompute.com/products/nvidia-b300 — B300 specs
- cudocompute.com/products/nvidia-gb200-nvl72 — GB200 specs
- cudocompute.com/products/nvidia-gb300 — GB300 specs
- cudocompute.com/products/nvidia-vera-rubin — Vera Rubin specs
- cudocompute.com/products/nvidia-b100 — B100 specs
- cudocompute.com/products/clusters — Cluster scale 128 to 40,000 GPUs
- cudocompute.com/service-level-agreement — 99.5% uptime for committed resources, service credits
- cudocompute.com/terms-of-service — Contract documents, DPA, English law
- rest.compute.cudo.org/v1/vms/machine-types-2 — VM GPU, vCPU, RAM, disk and IPv4 prices
- rest.compute.cudo.org/v1/machines-types — Bare-metal node prices and specs
- rest.compute.cudo.org/v1/clusters/machine-types — Cluster node price, InfiniBand
- rest.compute.cudo.org/v1/vms/gpu-models — GPU catalogue and VRAM
- rest.compute.cudo.org/v1/vms/data-centers — Regions, disk, object storage, network and IPv4 prices, supplier name
- docs.cudocompute.com/guides/billing-accounts/index — Prepaid credit, per-second precision, deletion at zero balance
- docs.cudocompute.com/guides/billing-accounts/create — Billing details required, Stripe card storage
- docs.cudocompute.com/guides/billing-accounts/top-up — Top-up flow
- docs.cudocompute.com/help/billing/what-payment-methods-are-supported — Cards, PayPal, large payments via ticket
- docs.cudocompute.com/help/general/what-is-the-minimum-top-up-value — 10 USD minimum, 10 to 1000 USD range
- docs.cudocompute.com/help/general/am-i-charged-if-a-vm-is-stopped-powered-off — Stopped VMs keep billing
- docs.cudocompute.com/help/billing/what-is-the-difference-between-hourly-and-monthly-costs — Monthly display uses 730 h
- docs.cudocompute.com/help/billing/when-do-i-get-invoiced-for-my-use — Monthly invoices and thresholds
- docs.cudocompute.com/help/accounts/how-can-i-create-an-account — Sign-up methods, email verification
- docs.cudocompute.com/help/general/how-can-i-get-support-for-my-problem — Console support tickets
- docs.cudocompute.com/help/general/where-can-i-check-the-service-status — Status page exists
- docs.cudocompute.com/guides/machines/index — Bare metal up to 8 GPUs
- docs.cudocompute.com/guides/clusters/index — Self-serve clusters, InfiniBand, NFS volumes
- docs.cudocompute.com/guides/virtual-machines/index — VM features, OS images
- docs.cudocompute.com/tools/terraform — Official provider CudoVentures/cudo v0.11.3
- docs.cudocompute.com/tools/cudoctl — Official CLI
- docs.cudocompute.com/tools/python-client — Python SDK
- docs.cudocompute.com/apps/index — Quick-deploy apps
Questions about CUDO Compute
What GPUs does CUDO Compute offer?
CUDO Compute has 4 GPU types with a fresh offer today, for example H100 from $1.82, L40S from $0.88 per GPU-hour on-demand. The full list with prices and market medians is on this page.
How does CUDO Compute bill?
Billing granularity: per second. Stopped instance: All billable resources keep accruing charges while a VM is stopped or powered off; resources must be deleted to stop billing
How trustworthy is CUDO Compute?
Its Trust Score is 80/100 (v0.1), read from its own documents. Strong on infrastructure control, security & compliance. Weaker on company & track record. The score measures transparency and published guarantees, not customer satisfaction; every criterion and its source is listed above.
Where are CUDO Compute's data centres?
2 named regions in 2 countries: Norway, Sweden.
How do you pay CUDO Compute, and are there credits?
Payment by card, crypto. No sign-up credit is published.
