Dec 1, live webinar: Agents in production, live: who gives the agent production access?

Sopra Steria and Scaleway's sovereign cloud deal: enough to leave US hyperscalers, or do you still need your own deployment platform?

Sopra Steria and Scaleway announced a sovereign cloud partnership on 19 November 2026. Here is what it covers, what it does not (SecNumCloud is still in progress on Scaleway's side), and the deployment, environment, audit and portability layer you still own after the migration project ends.

Romaric Philogene
CEO & Co-founder
NOV 22, 2026 · 17 MIN
Sopra Steria and Scaleway's sovereign cloud deal: enough to leave US hyperscalers, or do you still need your own deployment platform?

Key points:

  • The deal is real and recent. On 19 November 2026, Sopra Steria and Scaleway announced a strategic partnership to migrate critical applications and data to a sovereign environment, aimed at defence, aerospace, space, financial services and the public sector, in France, Belgium, Italy, Poland and the Netherlands first, with Germany and Sweden next (Sopra Steria press release).
  • It is not a SecNumCloud-qualified offer today. The announcement never mentions SecNumCloud, and Scaleway's own qualification is still "en cours" with ANSSI. If a tender names SecNumCloud for that exact workload, this partnership alone does not answer it yet.
  • It gives you infrastructure and a services contract. Scaleway brings European cloud and AI infrastructure; Sopra Steria brings consulting, integration, migration and cybersecurity. Together they cover jurisdiction, capacity and the migration project.
  • You still own the delivery layer. After the integrator's project ends, someone has to deploy every change, run environments, keep CI/CD working, produce audit evidence and keep the exit option open between Scaleway and any other cloud. That is a deployment platform problem, and no IaaS contract solves it.
  • Qovery covers that layer on Scaleway and everywhere else. Qovery, the Agentic DevOps platform, deploys your apps on Scaleway Kapsule in your own Scaleway project (Qovery-managed, or your existing cluster via the Qovery Operator), pushes images to Scaleway Container Registry, and gives you the same git-push, environment, RBAC and audit log workflow on AWS, GCP or Azure. Your apps never run on Qovery's infrastructure.

Qovery · Infrastructure control plane
Kubernetes, operated through one governed API
Learn more

When a European company tells me it is "moving to sovereign cloud", my first question is who will deploy the 40th release after go-live. The answer is usually a silence. The infrastructure contract is signed. The integrator has a migration plan. Nobody owns the delivery workflow that the hyperscaler's ecosystem used to provide for free.

The Sopra Steria and Scaleway partnership, announced on 19 November 2026, is a good answer to the first two questions. Europe is spending heavily here: Gartner forecasts worldwide sovereign cloud IaaS spending of $80 billion in 2026, up 35.6%, with Europe growing 83% from $6.9 billion in 2025 (Gartner, via The Register). Gartner also expects European spending to more than triple between 2025 and 2027 (DatacenterDynamics). So the question in the title is worth answering carefully. I run Qovery, and we deploy on Scaleway every day, so I will be explicit about where we fit and where we do not.

What did Sopra Steria and Scaleway announce on 19 November 2026?

Sopra Steria and Scaleway signed a strategic partnership to help European organisations migrate their most sensitive applications and data to a sovereign environment, from migration strategy through to running the applications. Sopra Steria brings consulting, integration, critical systems migration and cybersecurity; Scaleway brings cloud and AI infrastructure "designed and operated in Europe and built on open technologies" (press release, 19 November 2026).

Here is what the announcement says, and what it leaves out:

ItemWhat the 19 November 2026 announcement says
PartiesSopra Steria (consulting, integration, migration, cybersecurity) and Scaleway, the cloud and AI subsidiary of the iliad Group
GoalMove critical applications and data to a sovereign environment under European jurisdiction, with business continuity and "the ability to evolve or reverse their technology choices"
Target sectorsAeronautics, defence, security and space, financial services, public sector
CountriesFrance, Belgium, Italy, Poland, the Netherlands, and soon Germany and Sweden
Regulatory drivers namedNIS2, and DORA for the financial sector
AIGenerative AI use cases on confidential data, on Scaleway compute, with Sopra Steria integrating the AI solutions
PositioningStrengthens the "trusted cloud capabilities" Sopra Steria already offers its clients
Not mentionedSecNumCloud or any other qualification, pricing, named customers, service catalogue, deployment tooling

Two pieces of context help. Sopra Steria already works with other sovereign stacks: in February 2026 it announced it would integrate OVHcloud's On-Prem Cloud Platform for air-gapped and on-site deployments (Sopra Steria). So Scaleway joins other sovereign options in Sopra Steria's portfolio. And Scaleway has momentum with exactly these sectors: in July 2026 Airbus signed a two-year deal to move business-critical workloads to Scaleway (ITPro).

I like this partnership. A French integrator with deep critical-systems experience plus a European cloud with real Kubernetes and GPU capacity is the combination most regulated buyers have been asking for. My caution is about what buyers will assume it includes.

Is the Sopra Steria and Scaleway offer SecNumCloud-qualified?

No source I could find says so. The press release does not mention SecNumCloud at all, and Scaleway's own SecNumCloud qualification is still in progress with ANSSI, so you should treat the joint offer as "sovereign by jurisdiction and operations" and not as "SecNumCloud-qualified" until ANSSI lists Scaleway as qualified.

The timeline, from primary and trade sources:

  • January 2025. Scaleway entered the ANSSI process and passed the J0 milestone "without any reservations", targeting qualification by the end of 2025 (Next, Scaleway).
  • End of 2025. That target passed without a qualification.
  • August 2026. Scaleway described its initial scope: Instances, GPU, Object and Block Storage, VPC, Load Balancer, IAM, Key Manager, Audit Trail and Elastic Metal, in a single Paris availability zone, with Kapsule on the later PaaS roadmap (Scaleway).
  • November 2026. ANSSI's SecNumCloud page and Scaleway's security page still describe it as "en cours de qualification" as of our October check, and I found no announcement of a completed qualification since.

Why it matters: France's "Cloud au centre" doctrine and the SREN law push State data of particular sensitivity onto SecNumCloud-qualified offers. For those tenders, "in progress" is a no. We covered which providers are qualified today, and for which services, in our Numspot, OVHcloud and Scaleway SecNumCloud comparison.

For everything else (NIS2 entities, DORA-regulated firms, defence suppliers whose contracts ask for EU jurisdiction and EU operations but do not name SecNumCloud), Scaleway is already a credible answer. It is a French company in the iliad Group, outside the reach of the US CLOUD Act, which lets US authorities compel US providers to hand over data "regardless of whether such communication, record, or other information is located within or outside of the United States" (18 U.S.C. § 2713).

Is a sovereign IaaS plus an integrator enough to move sensitive workloads off US hyperscalers?

It is enough to move them, and too little to keep running them well. A sovereign IaaS answers where the workload runs and under which law; an integrator answers how you get there. Neither answers how your engineers ship change number 500 a year later, or how you would leave Scaleway if you had to.

Here is how I split the responsibilities. The middle column is my reading of the announcement, not a contract.

LayerCovered by Scaleway + Sopra Steria?What you still own
Jurisdiction and data locationYes: European provider, EU data centresChoosing regions, key custody, sub-processor review
Compute, storage, network, GPUsYes: ScalewaySizing, quotas, capacity planning
Migration strategy and executionYes: Sopra SteriaAcceptance criteria, cutover decisions
Security architecture and hardeningYes, as a service: Sopra SteriaDay-to-day patching and access reviews after handover
Kubernetes clustersPartly: Scaleway Kapsule is a managed serviceUpgrades, node pools, add-ons, policies
Deployment workflow (build, release, rollback)Not describedWho can deploy what, how, and how fast
Environments (dev, staging, previews)Not describedCreating, cloning, stopping and paying for them
CI/CD integrationNot describedPipelines, secrets, image registry, promotion
Audit evidence for deploymentsPartly: Scaleway Audit Trail covers cloud API callsWho deployed which version, when, with what approval
Portability and exitPromised as a goal ("evolve or reverse")A tested way to run the same apps on another provider

The last four rows are where hyperscaler migrations quietly fail. On AWS, teams lean on CodePipeline, ECS deployment circuit breakers, IAM-integrated CI, CloudTrail and a dozen marketplace tools without noticing. On a European IaaS, the primitives are there (Kapsule, Container Registry, Audit Trail, Secret Manager), but the glue is yours. An integrator can build that glue during the project. The question is who maintains it once the project team rolls off.

In many migrations, the integrator writes a careful set of GitLab CI templates, Helm charts and Terraform modules, hands them over with good documentation, and eighteen months later three people understand them. That is a key-person risk, and regulators have started asking about it.

What do you still need after the migration project ends?

You need five capabilities that outlive the integrator's statement of work: a repeatable deployment path, managed environments, CI/CD that targets the new cloud, deployment-level audit evidence, and portability you have tested. Each one is small on its own. Together they are the reason platform teams exist.

A deployment path developers can use without a ticket

Developers should push code and get a running, observable deployment on Kapsule without opening a ticket with the platform team or the integrator. That means a build step, an image in a registry you control (Scaleway Container Registry, in your own project), a Kubernetes rollout with health checks, and a one-click rollback. If every release still goes through a change request to the integrator, you have moved your infrastructure but not your speed.

Environments that match production and can be switched off

Regulated teams need staging environments that mirror production closely enough for testing to mean something, plus short-lived environments for each pull request. They also need to stop them at night. Non-production environments that run 24/7 are one of the most common sources of cloud waste, and GPU environments make it worse.

CI/CD wired to the new provider

Your CI (GitLab, GitHub Actions or something else) has to authenticate to Scaleway, push images, and promote releases. On a hyperscaler, many teams relied on native OIDC federation and managed deploy services. On Scaleway, you design that wiring yourself, store the credentials in a secrets manager, and keep it working through every rotation.

Audit evidence at the deployment level

Scaleway's Audit Trail records calls to Scaleway APIs. Your auditor will also ask who deployed version 4.2.1 of the payments service, when, from which commit, and who approved it. DORA applies to EU financial entities from 17 January 2025 and expects ICT change management and incident evidence (Regulation (EU) 2022/2554). NIS2 sets similar risk management obligations for essential and important entities (Directive (EU) 2022/2555). Both need deployment logs that tie a change to a person or an agent.

Portability you have tested

DORA Article 28(8) requires financial entities to have exit strategies for ICT services that support critical or important functions (EUR-Lex). The press release itself promises "the ability to evolve or reverse their technology choices". Promises are cheap; a quarterly test where you redeploy one real service on a second provider is what makes them true. Melanie wrote a full exit checklist for sovereign cloud providers that I recommend before you sign anything.

Keep shipping the same way on a sovereign cloud.
Qovery, the Agentic DevOps platform, deploys and operates your apps on Kubernetes in your own Scaleway, AWS, GCP or Azure account, or on a cluster you already run. Talk to us about your sovereign migration.

Which options do you have for the deployment layer on Scaleway?

You have five realistic options: let the integrator build and run a custom toolchain, assemble an open source GitOps stack yourself, use Scaleway's serverless products, stay on a hyperscaler's sovereign variant, or put a deployment platform such as Qovery above Kubernetes. Which one is right depends on how many teams deploy, how much in-house Kubernetes skill you have, and how seriously you take the exit clause.

ApproachWhat it isBest forMain trade-offPortability across clouds
Integrator-built toolchain (Sopra Steria or other)GitLab CI, Helm, Terraform and runbooks built during the project, often run as a managed serviceOrganisations that want one throat to choke and have budget for ongoing managed servicesChange velocity depends on the services contract; knowledge sits with the integratorGood if the toolchain is Kubernetes and Helm based; depends on discipline
DIY GitOps (Argo CD or Flux, Helm, Terraform, Backstage)Open source stack run by your own platform teamTeams with 3+ experienced platform engineersYou build and maintain every piece, including previews and auditHigh, if you avoid provider-specific services
Scaleway serverless (Serverless Containers, Functions)Scaleway-managed runtime for containers and functionsSimple stateless services, small teamsProvider-specific; less control than Kubernetes for complex systemsLow to medium
Hyperscaler sovereign variant (AWS European Sovereign Cloud, Bleu, S3NS)Hyperscaler technology run by an EU entityTeams deep in AWS, Azure or GCP tooling who mostly need EU operationsStill depends on a US vendor's software roadmapLow, if you use proprietary services
Qovery on Scaleway KapsuleQovery, the Agentic DevOps platform, deploying on Kapsule in your own Scaleway projectProduct teams without a large platform team that want the same workflow on Scaleway and a hyperscalerQovery's control plane is an extra vendor to review; it is not SecNumCloud-qualified itselfHigh: same workflow on Scaleway, AWS, GCP, Azure or any conformant cluster

The AWS European Sovereign Cloud became generally available in January 2026 in Brandenburg, Germany (Amazon), and it is a serious option if you mostly need EU operations and EU staff. Its technology and roadmap still come from AWS, which makes it a different answer from Scaleway. We compared Bleu, S3NS and Scaleway for SaaS scale-ups in more depth.

A combination is common and sensible: Sopra Steria leads the migration and security architecture, and the product teams deploy through a platform they control. The integrator's work then lands in something your own engineers will still use in three years.

How do you keep the exit option the partnership promises?

Keep your workloads in containers on standard Kubernetes, keep provider-specific managed services out of the critical path where you can, describe infrastructure as code, and keep the deployment workflow above the provider. Then test the exit on a schedule. Those four habits turn "the ability to reverse technology choices" from a slide into evidence.

A practical checklist for a Scaleway migration:

  1. Package everything as OCI images and push them to Scaleway Container Registry in your own project, with a copy policy to a second registry for critical services.
  2. Use Kapsule for orchestration and stick to upstream Kubernetes APIs. Kapsule runs upstream Kubernetes, so the same manifests and Helm charts run on EKS, GKE, AKS or a self-managed cluster.
  3. Prefer portable data services. Scaleway Managed PostgreSQL and S3-compatible Object Storage are easy to mirror elsewhere. Anything proprietary needs a written exit path.
  4. Keep Terraform per provider, with the application layer described once.
  5. Separate the deployment workflow from the provider, so moving a service means pointing the same environment at a different cluster.
  6. Run an exit drill at least twice a year: redeploy one real service on another provider, restore its data and measure the time. Our seven sovereignty tests give you the scenarios.

Migration scopes rarely include point 5, yet it decides whether point 6 takes a day or a quarter.

What should a migration plan with Sopra Steria and Scaleway look like?

Split the plan into three tracks that run in parallel: the infrastructure and security track (integrator plus Scaleway), the application delivery track (your platform owner), and the compliance evidence track (risk and audit). Most failed sovereign migrations I have looked at ran only the first one.

PhaseInfrastructure and security (Sopra Steria + Scaleway)Application delivery (you)Compliance evidence (risk, audit)
1. AssessWorkload inventory, data classification, target architectureDecide the deployment platform and environment modelMap NIS2, DORA or tender requirements to controls
2. Build landing zoneScaleway projects, IAM, VPC, Kapsule clusters, Key ManagerConnect clusters to the deployment platform, set up registries and secretsConfirm SecNumCloud status for each service in scope
3. Migrate in wavesData transfer, cutover runbooks, security testingRecreate environments, wire CI, run preview environments for each changeCollect deployment and access logs per wave
4. RunManaged services, security operations (if contracted)Teams deploy on their own, environments stop when idleQuarterly access reviews, exit drill evidence
5. Prove the exitSecond-provider landing zone (Scaleway plus one more)Redeploy a critical service on the second providerDocument the measured exit time

Start the delivery track in phase 1, not phase 4. If you pick the deployment platform after the clusters exist, you will rebuild half of phase 2.

How does Qovery handle deployment on Scaleway and other sovereign clouds?

Qovery, the Agentic DevOps platform, is the delivery layer in the table above. It deploys and operates your applications on Kubernetes in your own Scaleway project, so the clusters, the data, the images and the Scaleway bill stay yours. Qovery never hosts your applications.

What you set up on Scaleway:

  • A Kapsule cluster. Qovery can create and manage a Kapsule cluster in your Scaleway project, or you connect an existing Kapsule cluster with the Qovery Operator (BYOK v2). The Operator installs once with Helm, Qovery never stores your kubeconfig, the Operator opens an outbound connection, and images are built inside your cluster and pushed to Scaleway Container Registry in your account (Qovery BYOK docs).
  • Environments. Git-push deployments, a preview environment per pull request, environment cloning, and stopping non-production environments when nobody uses them.
  • Controls. RBAC, audit logs of every deployment and configuration change, secrets manager integrations, Helm and Terraform or OpenTofu services, and a Terraform provider so the setup is code.
  • The same workflow elsewhere. The same environments run on AWS, GCP or Azure, or on any conformant cluster you connect with qovery cluster install (OVHcloud, Outscale, on-premise). Cross-cloud networking is not built in; configure it with a site-to-site VPN, Cilium Cluster Mesh or Tailscale.

What Qovery does not do: it is not SecNumCloud-qualified, it does not make you compliant, and it does not replace an integrator for data migration or security architecture. It helps you meet controls (audit logs, RBAC, region choice, your own accounts).

How to start: the Free plan includes 3 users, one cluster of your own and 1,000 deployment minutes a month, with no credit card. Run qovery demo up to try it on a local cluster first, or connect your Kapsule cluster with the Operator.

How do I set this up from Claude Code or Cursor?

Install the Qovery Agent Skills, then ask your coding agent in plain language to set up Qovery on your Scaleway project and deploy your app. The agent proposes a plan and you approve each action; the MCP Server stays read-only unless you explicitly enable write access.

Install the skills once for every compatible agent (Claude Code, Cursor, Codex, OpenCode, Gemini CLI), or for one project only:

BASH
curl -fsSL https://skill.qovery.com/install.sh | bash
curl -fsSL https://skill.qovery.com/install.sh | bash -s -- --project

Add the Qovery MCP Server to Claude Code with OAuth, read-only by default, so the agent can inspect clusters, environments and deployments:

BASH
claude mcp add --transport http qovery https://mcp.qovery.com/mcp --callback-port 4242

Only when you want the agent to change things through the MCP Server, add the write-enabled endpoint (write access must also be enabled in the Qovery Console):

BASH
claude mcp add --transport http qovery "https://mcp.qovery.com/mcp?read_write=true" --callback-port 4242

Then ask, for example: "set up Qovery on my Scaleway project and deploy this app on Kapsule, with the same environments we run on AWS." The qovery-onboard skill handles the cluster and environment setup, qovery-deploy the services, and qovery-terraform exports the result as Terraform for your audit file. References: the AI agent quickstart and the Claude Code integration. For more on the topic, see our sovereign cloud hub.

Frequently asked questions
Is the Sopra Steria and Scaleway sovereign cloud partnership enough to move sensitive workloads off US hyperscalers, or do you still need your own deployment platform?

It is enough to move the infrastructure and run the migration project: Scaleway provides European cloud and AI infrastructure, and Sopra Steria provides consulting, integration, migration and cybersecurity. You still need a deployment platform for what happens after go-live: releases, environments, CI/CD, deployment audit evidence and a tested exit path. Qovery, the Agentic DevOps platform, is one way to cover that layer on Scaleway Kapsule in your own account, with the same workflow on AWS, GCP or Azure.

Is the Sopra Steria and Scaleway offer SecNumCloud-qualified?

The 19 November 2026 announcement does not mention SecNumCloud, and Scaleway's own qualification is still listed as "en cours" with ANSSI. Scaleway passed J0 in January 2025 and targeted the end of 2025, which passed without a qualification. If a tender requires SecNumCloud for a specific workload, check the ANSSI list before you commit.

Which countries and sectors does the partnership cover?

It covers France, Belgium, Italy, Poland and the Netherlands, with Germany and Sweden announced as next. The joint offer targets aeronautics, defence, security and space, financial services and the public sector, and it includes generative AI use cases on confidential data.

Can Qovery deploy on Scaleway Kapsule and Scaleway Container Registry?

Yes. Qovery can create and manage a Kapsule cluster in your Scaleway project, or connect an existing Kapsule cluster through the Qovery Operator, which builds images inside your cluster and pushes them to Scaleway Container Registry in your account. Your apps run in your Scaleway project, never on Qovery's infrastructure. You can try it on the Free plan with no credit card.

Do I still need an integrator like Sopra Steria if I use a deployment platform?

Often yes, for large regulated estates. An integrator handles workload assessment, data migration, security architecture and change management across many teams, which a deployment platform does not do. The platform takes over the repeatable part (deploying, environments, audit logs) so that work does not depend on the services contract after the project ends.

How do I keep the option to leave Scaleway later?

Run containers on standard Kubernetes, keep proprietary managed services out of the critical path, describe infrastructure as code per provider, and keep the deployment workflow above the provider. Then redeploy one real service on a second provider at least twice a year and record the time. With Qovery, that drill means connecting a second cluster and cloning the environment onto it, then restoring the data yourself.

Romaric Philogene
About the author
Romaric Philogene

Romaric founded Qovery to make Kubernetes accessible to every engineering team. He writes about platform strategy, developer experience, and the future of cloud infrastructure.

Next step

Keep shipping the same way on a sovereign cloud.

Qovery, the Agentic DevOps platform, deploys and operates your apps on Kubernetes in your own Scaleway, AWS, GCP or Azure account, or on a cluster you already run. Talk to us about your sovereign migration.