Q1 2026 - Quarterly Product Update


The Cloud Cover

Q1 2026 Product Update

Welcome to The Cloud Cover, your quarterly look at what’s shaping the OVHcloud platform.

Vimeo conditions the playback of its videos on the deposit of tracers in order to offer you targeted advertising based on your browsing.

In order to watch the video, you need to accept the Sharing cookies on third-party platforms privacy category in our Privacy Center. You have the option of withdrawing your consent at any time.

For more information,visit the Vimeo cookies policy and the OVHcloud cookies policy .

Platform Foundations & Resilience

Everything starts with the foundation.

Because today, if you're running production workloads downtime isn’t an option.
Unpredictability isn’t acceptable.Resilience has to be built in from day one.That’s exactly what we’re doing.
 

Managed Kubernetes Multi AZ Resilience · GA · September 2025

Managed Kubernetes Standard - Enteprise grade, finally

After years of a Free-only tier, our Managed Kubernetes Service now has a Standard plan built for production. A 3-AZ control plane, dedicated etcd up to 8 GB, and clusters scaling to 500 nodes. This is Kubernetes the way enterprises actually need it.

99.99%

SLA

500

Max nodes

8 GB

Dedicated etcd

Landing Zone · GA · February 2026

Landing Zone — Your secure foundation before moving to the cloud.

Deploying a Landing Zone is a best practice to enhance security and governance. We now offer pre-configured "base camps" for Small Business (OPNSense, basic IAM, shared vRack), Enterprise (Fortinet/Stormshield HA, multi-vRack, SSO federation), and Government (CCCS/Protected-B compliance, IP ACL, data residency).

3

Tiers available

from SMBs to Gov

Ready for all organizations

Public Cloud Instances · Observability Metrics · coming soon

Observability Metrics — Native Prometheus monitoring for OVHcloud.

Cloud-native, fully Prometheus-compatible managed service for collecting, storing, and querying time-series metrics. First integration: Instances Metrics Dashboard. Automatic visibility on CPU, RAM, network, and performance without complex setup. 100% compatible with remote write, PromQL, and alerting. Currently in Gravelines only. The first milestone of unified OVHcloud Observability.

PromQL

Compatible

Free

With instances

Milan Datacenter · GA · November 2025

Milan 3-AZ, second Datacenter of our new Standard

Customers can deploy workloads across availability zones while keeping applications within the same geographic region.
Multi-AZ is becoming our new standard — and more regions are already on the way

vRack · RTvRack · January 2026

RTvRack — Public IP routing into your private network.

Expose services hosted inside private vRack networks to the public Internet using Additional IP blocks. New bandwidth options now available in EU/CA/APAC beyond free tiers. Scale outbound capacity on demand with Anti-DDoS and Edge Firewall protection. No complex NAT or proprietary networking — just direct public IP routing into a global private network.

5 Gbps

EU/CA free tier

DDoS Protection

Included

FinOps · FOCUS compatibility · February 2026

FinOps — FOCUS 1.0 compliant data for Public Cloud.

Take control of your cloud costs with standardized, actionable insights adhering to the FOCUS (Financial Operations) standard. Download FOCUS-compliant data and integrate with third-party platforms (Lota.cloud, Holori) or your own data warehouses. Gain visibility into usage and costs, identify optimization opportunities, and improve financial governance across finance, engineering, and operations teams.

FOCUS

1.0 compliant

All

Public Cloud services

Border Gateway Protocol · Alpha · February 2026

BGP as a Service — Dynamic routing for advanced network architectures.

Border Gateway Protocol now available via OneAPI. Configure BGP peerings with IPv4 and IPv6 support across subnets and networks. Alpha 2 now open on 11 availability zones (EU, CA) with 4 active clients. First consumer: Public Cloud in Milan 3-AZ. Next: Public VCF-as-a-Service and vRack Services. Beta phase scheduled for end Q3 FY26.

11

AZ available

Alpha

Now open

Data, Storage & AI Acceleration

More data. More workloads. More AI.
And that changes what the platform needs to support.

 

Cloud GPUs NVIDIA H200 · GA · March 2026

GPU power

H200 instances now available in multiple regions, delivering the performance required for the most demanding AI and high-performance computing workloads.

Managed ClickHouse · GA · February 2026

ClickHouse — Billions of rows. Milliseconds of response.

No hyperscaler offers the original ClickHouse in a managed solution. We do. Fully managed, multi-AZ, and purpose-built for real-time analytics: dashboards, IoT telemetry, AdTech pipelines, user behaviour analysis. A natural cross-sell for any DBaaS customer evolving toward analytical workloads — and a sovereign alternative to self-hosted complexity.

3-AZ

Exclusive launch

ms

Query at scale

Data Platform · ISO 27001 & HDS · February 2026

Data Platform — Now ISO 27001 and HDS certified.

The OVHcloud Data Platform, designed to accelerate your Data & Analytics projects, has obtained ISO 27001 and HDS certifications. These certifications attest to the high level of security, regulatory compliance, and trustworthiness. Essential for addressing regulated customers, public sector organizations, and healthcare players.

ISO

27 001 certified

HDS

Healthcare ready

AI Endpoints - LLM Guards · Beta · February 2026

LLM Guard — Secure your LLM inputs and outputs.

A control layer that analyzes and filters prompts (inputs) and generated responses (outputs). Limit sensitive or inappropriate content — policy violations, dangerous instructions, misuse — and strengthen control of production deployments. Currently in free beta. An essential building block for secure LLM applications.

Free

Beta acces

Safe

Content filtering

End-to-End Encryption · Acquisition · December 2025

OVHcloud acquires SEALD — Encryption becomes a native service.

The average cost of a data breach is now $4.4M. The encryption market is growing 16.2% annually. Our acquisition of SEALD brings certified, developer-friendly E2EE directly into the Public Cloud universe — so developers can encrypt client-side without managing cryptographic complexity. Encryption is trust.

AI Deploy new feature · GA · February 2026

Autoscaling down to 0

Bi-directional autoscaling now lets you automatically scale deployments to 0 GPU instances during inactivity, then restart on the first request. Dramatic cost reduction for workloads with business cycles — working hours, test environments, batch jobs, POCs. Pay only for what you actually use.

AI Endpoints - Vision Language Model · GA · February 2026

VLM — Payload size increased to 10 MB.

Good news for Vision-Language Models: maximum payload size jumps from 2 MB to 10 MB. Send larger, higher-resolution images without excessive compression. Better support for complex documents, detailed visuals, industrial captures — and less pre-processing on the application side.

10 MB

Max Payload

5X

Increase

Object Storage - Active Archive · GA · February 2026

Active Archive — Long-term storage, instant access.

We now offer five Object Storage classes — covering every workload from blazing fast SSD to deep cold archive. The new Active Archive class fills the gap between infrequent access and cold storage: 99.99% SLA, 3-AZ resilience, and instant retrieval for compliance-critical data at ~457.71 ₹/TiB/month.

99.99%

SLA

5

Storage Classes

457.71 ₹

Per TiB/month

Enterprise File Storage · GA · February 2026

EFS — NetApp power, sovereign delivery.

50% of File Storage users run Kubernetes workloads. EFS now supports the Trident CSI driver for dynamic NFS provisioning — bringing guaranteed 4,000 IOPS/TiB and 99.99% SLA to stateful Kubernetes at enterprise scale. Up to 2.5× cheaper than comparable products. Fully immune to the US Cloud Act.

4k

IOPS per TiB

2.5x

Cost Advantage

99.99%

SLA

Quantum - Emulators & QPUs · GA · February 2026

Quantum computing — Emulators and QPU access expanding.

New emulators available: Pennylane, Tii, and Qunasys. Plus QPU access via Quandela Belenos. OVHcloud continues to invest in quantum readiness through partnerships with Teratec, Lab quantique meetups, and industry events across Europe.

3

New Emulators

QPU

Quandela Belenos

AI Endpoints - Vision Language Model · GA · February 2026

VLM — Payload size increased to 10 MB.

Good news for Vision-Language Models: maximum payload size jumps from 2 MB to 10 MB. Send larger, higher-resolution images without excessive compression. Better support for complex documents, detailed visuals, industrial captures — and less pre-processing on the application side.

10 MB

Max Payload

5X

Increase

AI Endpoints - LLM Guards · Beta · February 2026

LLM Guard — Secure your LLM inputs and outputs.

A control layer that analyzes and filters prompts (inputs) and generated responses (outputs). Limit sensitive or inappropriate content — policy violations, dangerous instructions, misuse — and strengthen control of production deployments. Currently in free beta. An essential building block for secure LLM applications.

Free

Beta acces

Safe

Content filtering

Quantum - Emulators & QPUs · GA · February 2026

Quantum computing — Emulators and QPU access expanding.

New emulators available: Pennylane, Tii, and Qunasys. Plus QPU access via Quandela Belenos. OVHcloud continues to invest in quantum readiness through partnerships with Teratec, Lab quantique meetups, and industry events across Europe.

3

New Emulators

QPU

Quandela Belenos

Public Cloud File Storage · Beta · January 2026

Public Cloud File Storage — Our second most-wanted product, delivered.

File Storage has been the #2 most requested Public Cloud feature for years — representing €33M in lost ARR since 2020. The Beta is now live, free, and open to all in GRA, DE1 and SBG5. Performance that beats GCP and Azure in the 150 GiB–1 TiB range. Pricing 5% below AWS. Full data sovereignty.

16k

Max IOPS

2.5x

Free Beta access

Cloud GPUs NVIDIA H200 · GA · March 2026

GPU power

H200 instances now available in Paris region, delivering the performance required for the most demanding AI and high-performance computing workloads.

8

Hosts available

8

H200 GPU/host

141

GB of HBM3e Memory

AI Deploy new feature · GA · February 2026

Autoscaling down to 0

Bi-directional autoscaling now lets you automatically scale deployments to 0 GPU instances during inactivity, then restart on the first request. Dramatic cost reduction for workloads with business cycles — working hours, test environments, batch jobs, POCs. Pay only for what you actually use.

Hybrid cloud & Enterprise platforms

Cloud transformation isn’t starting from scratch.
It’s about evolving what already exists securely, progressively, and without compromise.

 

Developer & SMB Experience

The platform easier and faster to build on.