Provider record · updated 2026-09-10
OVHcloud — GPU pricing & deliverability
4 tracked offerings · grade mix A:0 · B:0 · C:1 · D:3 · U:0. Public Cloud GPU page (USD, Worldwide site) publishes 'Up to 25 Gbps included' for public and private networks and '1-4 GPUs per instance' as general specs; H200 is the exception at 8 GPUs/instance with NVLink. No InfiniBand or multi-node cluster fabric published on fetched pages, and no contiguous cluster product described. A100 no longer appears on the GPU lineup page (only H200, H100, L40S, L4, A10 $1.00/hr, V100S $0.88/hr, Quadro RTX 5000 $0.60/hr). No B200, MI300X. The main /public-cloud/prices/ price list is client-rendered and did not expose GPU rows to static fetch.
Tracked offerings
| GPU | SKU | $/GPU-hr | Terms | Grade | Conf | Fabric (as published) | Src |
|---|---|---|---|---|---|---|---|
| L4 | Nvidia L4 (24 GB GDDR6, PCIe 4.0) GRA11 (Gravelines, FR) | $1.00 | on-demand | D | publ | Up to 25 Gbps Ethernet (public/private network) | ↗vendor |
| L40S | Nvidia L40S (48 GB GDDR6, PCIe 4.0) GRA11 (Gravelines, FR) | $1.80 | on-demand | D | publ | Up to 25 Gbps Ethernet (public/private network) | ↗vendor |
| H100-PCIe | Nvidia H100 (80 GB HBM2e, PCIe 5.0) GRA11 (Gravelines, FR) | $2.99 | on-demand | D | publ | Up to 25 Gbps Ethernet (public/private network) | ↗vendor |
| H200-SXM | Nvidia H200 (141 GB HBM3e, NVLink, 8-GPU instance) | $6.20 | on-demand | C | infe | Up to 25 Gbps Ethernet (public/private network) | ↗vendor |
Grades measure what sellers publish — deliverability of the written claim, not measured performance. No fleet holds a measured (Verified) badge yet; that tier arrives with the verification harness. Source classes: API = provider's live pricing API · vendor = provider's published page · mkt API = marketplace API · 3rd-party = aggregator (provider publishes no price on its own pages) · quote = quote-gated or "starting at", not a walk-up price. Quote-class prices are shown in tables but excluded from headline ranges, spreads, and fair-price bands.
Terms & trust (as published, 2026-09-14)
| Uptime SLA | 99.99% monthly availability for Public Cloud Instances including GPU instances (GPU page: 'The SLA guarantees 99.99% monthly availability on GPU instances'; US legal SLA: Instances and Load Balancer 99.99%, Discovery 99.95%, Storage 99.90%, Managed Kubernetes 99.50% with no credits) — measured on: External connectivity — 'The Service will be considered Unavailable...if it has no external connectivity'; unavailability is deemed to commence only once the customer opens a support case; monthly measurement, 'commercially reasonable efforts' standard |
| SLA remedy | credits — 10% of monthly charge if availability between 95.00% and the commitment, 100% if below 95.00%; capped at 100% of monthly invoiced amount; 'sole and exclusive remedy'; claim within 60 days, credits issued within 45 days, expire after 12 months. Exclusions: scheduled maintenance, customer misuse, DDoS/hacking, third parties beyond OVH's border router, force majeure. |
| Capacity guarantee | No published GPU capacity guarantee. Savings Plans (12 months -15%, 36 months -30%, self-serve in control panel) are usage-billing commitments, not reservations — page never uses 'reserved' or 'guaranteed capacity', overage billed at standard rates, and US eligibility lists only B3/C3/R3 instances and MKS nodes (GPU instances not listed). |
| Egress fees | Included/free in most regions — inbound included everywhere; outbound public traffic 'included in the price of instances in all locations except the Asia-Pacific region (Singapore, Sydney, Mumbai)', where each project gets 1 TB/month outbound then per-GB charges apply |
| Billing granularity | per-hour (standard), optional monthly billing (irreversible switch; monthly estimated at 730 hrs); storage per GB/hour; no per-second or per-minute billing published |
| Self-serve signup | yes — immediate signup with US$200 free credit; GPU instances (L4/L40S/H100/H200 etc.) deployable from the control panel at published hourly rates; large multi-node GPU deployments may be quoted via sales |
| Security certifications | ISO/IEC 27001 (plus 27017/27018), ISO 27701, SOC 1/2/3 attestations, HDS (France), SecNumCloud (ANSSI, scoped to specific trusted-cloud offerings such as Hosted Private Cloud — not the whole Public Cloud GPU line), HIPAA compliance claim (US data centers), PCI DSS, CSA STAR Level 1; GPU product page lists 'Certifications: ISO27001, SOC, Health data' |
Notes: SLA gotcha: downtime officially starts only when the customer opens a support case, which shrinks measured unavailability. 99.99% GPU figure from the worldwide GPU marketing page plus US legal SLA; EU-entity terms are in region-specific PDFs not fetched. Compliance page does not map certifications to specific product lines.
Sources: ↗ ↗ ↗ ↗ ↗ · Fields marked "not published" are exactly that — the seller does not state them anywhere we could find. Absence is a finding.
Reliability record (public status feed, last 90 days)
50+ incidents on OVHcloud's own status page in the trailing 90 days, collected 2026-09-15. The feed caps at 50 items — the true count is higher. Self-reported by the provider — a floor on true incident count, not a ceiling; it still beats the war-story thread you were going to read instead. High counts often reflect granular per-rack reporting, not worse infrastructure: read the incidents, not just the number.
| Date | Incident | Impact | Duration |
|---|---|---|---|
| 2026-09-13 | [GRA1/3/9/11][Public Cloud] - Instances incident notification ↗ | major | 1.3 h |
| 2026-09-13 | [DE1][Storage] - Ceph volumes incident notification ↗ | minor | 2.4 h |
| 2026-09-10 | [SBG5/7][Compute - Instance] - Some Instances Incident Notification ↗ | major | 23.7 h |
| 2026-09-06 | [GRA11][Containers & Orchestration] - Managed Kubernetes Service Incident Notification ↗ | major | 1.6 h |
| 2026-09-03 | [GRA1/GRA11][Storage] - Block storage incident notification ↗ | minor | 3.4 h |
| 2026-09-03 | [GLOBAL][Databases] - MongoDB incident notification ↗ | major | 21.4 h |
| 2026-08-31 | [GRA5/GRA7/GRA9/GRA11][Compute - Instance] - Compute Instance Incident Notification ↗ | major | 1 h |
| 2026-08-28 | [GRA1][Compute - Instance] - Instance incident notification ↗ | none | 0 h |
| 2026-08-25 | [GRA5/6/7/8/9/11][Compute - Instance] - Instance incident notification ↗ | major | 0.8 h |
| 2026-08-25 | [GLOBAL][Identity, Security & Operations] - User Management Incident Notification ↗ | major | 2.4 h |
| 2026-08-25 | [GLOBAL][Data & Analytics] - Cloud Database Incident Notification ↗ | major | 39.7 h |
| 2026-08-23 | [GRA3][Compute - Instance] - Compute Instance Incident Notification ↗ | major | 0.2 h |
What OVHcloud does not publish
- H200 exact instance name and per-GPU list price
- regions per GPU type
- max_contiguous_gpus
- reserved/savings-plan pricing
- min_commit
- notice_period
- H100/L40S NVLink details
Unpublished fields are recorded as unknown and grade conservatively. Sellers improve their grades by disclosing — that is the point.
Weekly context on every provider
The Delivered Compute Report tracks how these prices and grades move.
Subscribe free