Market Context and Selection Criteria
The cloud market remains dominated by three American hyperscalers – AWS, Microsoft Azure, and Google Cloud – which together command more than 70% of global market share and the bulk of R&D spending. Facing them, European players such as OVHcloud and Scaleway occupy a strategic niche: sovereignty. Cloud cost remains, for the third year in a row according to Flexera (2025), the IT leaders’ primary concern, cited by 84% of them.
The central criterion that has emerged is sovereignty. The European Court of Justice’s Schrems II ruling (2020) reminded us that data hosted with American providers remains subject to the CLOUD Act, regardless of its location. During a Senate hearing in June 2025, a Microsoft representative acknowledged that data sovereignty in Europe cannot be guaranteed. This highlights the core of ANSSI’s SecNumCloud qualification, which hyperscalers cannot structurally achieve for their standard offerings.
This comparison selects five providers based on three criteria: market share and maturity, the level of sovereignty (determinant in France), and the fit with workload profiles. The aim is not to crown a winner—sovereignty and breadth of services often involve trade-offs—but to qualify the use cases where each option excels. Other sovereign players, such as Outscale (Dassault) or Clever Cloud, complete this landscape for the most sensitive needs.
Synthetic Comparison Table
| Provider | Profile | Sovereignty | Preferred Target |
| AWS | Hyperscaler, global leader | Not sovereign (CLOUD Act) | Innovation, broad catalog, data/AI projects |
| Microsoft Azure | Hyperscaler, Microsoft ecosystem | Not sovereign (“trusted” offerings) | Organizations leveraging Microsoft 365 |
| Google Cloud | Hyperscaler, data and AI | Not sovereign | Analytics, machine learning, BigQuery |
| OVHcloud | European champion, French | Sovereign (partial SecNumCloud) | Public sector, controlled costs |
| Scaleway | French cloud, for developers | Sovereign (SecNumCloud underway) | Startups, economical AI workloads |
Detailed Solutions Overview
AWS
As a global leader, AWS (Amazon Web Services) offers the widest service catalog and the most mature ecosystem. It is the reference for innovation, data and AI projects (SageMaker), and complex architectures. It has regions in France (Paris). Its downside: costs are often 30–50% higher than sovereign actors for equivalent workloads, a learning curve, and exposure to the CLOUD Act that can exclude the most sensitive data.
Microsoft Azure
Azure draws its strength from the integration with the Microsoft ecosystem (365, Entra ID, Power Platform). For organizations heavily equipped with Microsoft tools, it is the frictionless path. Microsoft has strengthened its footprint in France (availability zones in Paris and Marseille in 2025) and offers trusted cloud offerings operated with partners (Orange, Capgemini), though it does not achieve the sovereign legal status of an independent European player.
Google Cloud
Google Cloud (GCP) stands out for data and AI use cases, with services like BigQuery and Vertex AI particularly valued for analytics and machine learning. It acts as an accelerator for data-driven projects. Like its American peers, it remains subject to the CLOUD Act, which can exclude it from certain public tenders or regulated data calls.
OVHcloud
Founded in 1999 in Roubaix by Octave Klaba, OVHcloud is the leading European cloud provider and the national champion, with revenue surpassing one billion euros in the 2024-2025 period. It holds a significant share of the European sovereign segment and manufactures its own servers, a cost-efficiency lever. Some of its services are certified or in the process of SecNumCloud qualification. Natural target: the public sector, sovereignty-minded organizations, and those seeking a strong cost/performance ratio (20-40% cheaper than AWS on compute).
Scaleway
Scaleway, a French player within the Iliad group, positions itself as the cloud for developers and proclaims a high degree of sovereignty (European servers and teams, independent capitalization). The company is progressing through the ANSSI’s SecNumCloud qualification process (started early 2025, milestone J0), with a target qualification in the long term. With some of Europe’s most competitive pricing and economical GPU offerings for AI, it appeals to startups, developers, and AI workloads with budget discipline.
How to Choose Based on Your Profile
Choice hinges on balancing service richness, sovereignty, and cost. Here are some guidelines:
- Innovation and complex data/AI projects, without strong sovereignty constraints: AWS or Google Cloud, for the depth of their catalogs.
- Microsoft 365 business: Azure, for native integration and ecosystem coherence.
- Public sector, sensitive data, markets demanding SecNumCloud: OVHcloud, or Outscale for the highest levels.
- Startups, developers, affordable AI workloads: Scaleway, for its pricing and sovereignty.
- Hybrid approach: combine a hyperscaler for elasticity with a sovereign actor for critical data—often the most realistic solution.
A methodological point should guide brand choice: decisions should stem from a data classification and placement strategy, not from a priori preferences. A health record, an industrial secret, and a corporate website do not call for the same vendor. With the price gap between sovereign and hyperscalers narrowing, the cost argument alone is no longer sufficient to justify legal exposure for critical data.
Finally, whatever the provider, preserving reversibility remains essential: favor open standards and portable technologies (containers, Kubernetes) to avoid vendor lock-in. The right approach is to begin with your real requirements—sovereignty, performance, budget, existing ecosystem—then test the chosen solution within a controlled scope. In a shifting geopolitical context, the ability to evolve your architecture matters as much as the initial choice.
This content is published by Mentioned
