For the past two years, “EU cloud sovereignty” has mostly been an enterprise and government story — banks worried about DORA audits, public-sector tenders that exclude AWS, defense ministries that can’t touch anything with US CLOUD Act exposure. Startups have largely watched from the sidelines, on the assumption that sovereignty rules are a big-company problem and AWS isn’t going anywhere for the rest of us.
A new research report from Gart Solutions, a Kyiv-based DevOps and cloud infrastructure firm, makes the case that this assumption is about to get expensive. The report, “The EU Cloud Managed Services Gap: An AWS Capability Breakdown and Sovereign Stack Opportunity,” benchmarks AWS against seven European cloud providers — OVHcloud, Scaleway, Hetzner, Exoscale, IONOS Cloud, UpCloud, and the Swedish sovereign-cloud newcomer evroc — across the services that actually run a modern application: managed databases, Kubernetes, observability, serverless, messaging, AI infrastructure, and backup/DR.
Its conclusion, in short: the infrastructure gap between AWS and EU clouds is closing. The managed services gap is not. And for startups, that distinction is about to matter a lot more than it used to.
The numbers behind the gap
The report’s headline stat is blunt. AWS exposes a catalog of more than 200 managed services, identically, across every region — Frankfurt and Dublin included. The deepest EU catalog the researchers found, OVHcloud’s, covers around 45. Scaleway, generally considered the most developer-friendly EU provider, covers around 30. Hetzner — beloved by bootstrapped startups for its pricing — covers around 8, by design: it sells infrastructure, not platform.
Kubernetes is the one category where Europe is genuinely competitive: several EU providers now beat AWS EKS on cluster deploy time, and free control planes are standard. But managed databases are where the story turns. Most EU providers can match Amazon RDS — a managed PostgreSQL instance with backups and failover. None of them match Aurora: the distributed, auto-scaling, scale-to-zero database tier that a lot of modern SaaS architectures are quietly built around. According to the report, no EU provider has a credible roadmap to close that specific gap, because the underlying engineering — replicated, distributed storage — is a multi-year undertaking, not a feature release.
Observability, event-driven/serverless runtimes, and managed AI infrastructure tell a similar story: present at AWS depth, present in patches across the EU, but nowhere near full coverage.
Why this is a startup problem, not just an enterprise one
It’s tempting to read this as “interesting, but I’m not a bank.” Three things suggest otherwise.
Your customers are starting to ask. If you sell into fintech, healthtech, insurtech, or anything touching EU public-sector procurement, you’ve probably already had a prospect’s security questionnaire ask where data lives and whether a US company can be compelled to access it. DORA’s incident-reporting and third-party risk requirements went live for financial entities in January 2025; NIS2 enforcement is active across member states. Neither regulation directly targets a five-person startup — but both regulations bind your customers, who will increasingly push data-residency requirements down into their vendor contracts. A startup that can say “fully EU-hosted, no US sub-processor” closes deals a few weeks faster than one that can’t. That’s not theoretical compliance risk; it’s sales cycle.
The migration tax is real, and small teams feel it disproportionately. This is the part the Gart Solutions report makes uncomfortably concrete. If you move off AWS to satisfy a sovereignty requirement, you don’t just lose Aurora — you likely lose RDS Proxy, Performance Insights, Lambda, EventBridge, and whatever observability stack you’ve been getting for free from CloudWatch. Somebody has to rebuild that operating layer: Patroni for HA Postgres, Prometheus/Grafana/Loki for observability, Knative or OpenFaaS for serverless. At a 2,000-person bank, that’s a platform team’s quarter. At a 12-person startup, that’s the entire engineering org’s quarter — at the exact moment you should be shipping features, not running database failover drills.
A new category of provider is emerging to absorb exactly that tax. This is the more interesting half of the report. Gart Solutions frames the gap not as a wall but as a market: systems integrators and managed service providers who build the missing operating layer once and deliver it as a service on top of EU infrastructure, rather than leaving each company to rebuild it from scratch. The report cites the partnership between Detecon and evroc — where evroc supplies sovereign IaaS and a systems integrator supplies the entire managed layer on top — as an early template for what this looks like at enterprise scale. The same logic scales down: a startup moving to OVHcloud or Hetzner doesn’t have to choose between “self-manage everything” and “stay on AWS.” A third option — buy the managed layer from a specialist rather than build it — is forming in real time, priced more like a SaaS subscription than a platform-engineering hire.
What this actually means if you’re building in Europe right now
A few practical takeaways follow from the report, independent of whether you ever plan to leave AWS:
- Architect for portability now, even if you don’t move now. The cost of the managed-services gap is almost entirely a migration cost, not a steady-state cost. Startups that avoid deep, hard-to-replace dependencies on AWS-specific managed services (Aurora’s specific scaling behavior, deep Lambda/EventBridge coupling) keep optionality cheap. You don’t need to be cloud-agnostic on day one. You do want to know which of your dependencies would be expensive to unwind later.
- Sovereignty is becoming a sales asset before it’s a legal requirement. If your buyer persona includes EU banks, insurers, hospitals, or government bodies, “EU-hosted, EU-operated” is shifting from a nice-to-have line in a pitch deck to a checkbox that gates the deal. Worth knowing before a procurement process surfaces it as a blocker six months into a sales cycle.
- Watch the SI/MSP layer, not just the cloud providers. The interesting startups and vendors to track over the next 18 months may not be the EU hyperscalers themselves but the specialists building the managed layer on top of them — managed Postgres-on-EU-cloud, compliance-as-code subscriptions calibrated to DORA and NIS2, sovereign DR orchestration. That’s where the gap actually gets solved, and it’s a market several systems integrators (Gart Solutions among them) are now explicitly positioning for.
- Don’t assume “EU cloud” means “feature parity with AWS.” If you’re evaluating a move for compliance reasons, budget for the operating-layer gap explicitly. The report’s most useful contribution for a founder doing this evaluation is exactly this: a category-by-category map of what you’ll actually lose, provider by provider, rather than a vague sense that “European cloud is behind.”
The bigger picture
None of this means AWS is going away for European startups, and the report doesn’t claim otherwise. What it does claim — credibly, with a service-by-service comparison to back it up — is that the sovereignty conversation has been having the wrong argument. The question was never really “is EU infrastructure good enough.” Compute, storage, and networking from OVHcloud, Hetzner, or Scaleway are perfectly serviceable for most workloads today. The real question is who builds and operates everything sitting on top of that infrastructure — and right now, in Europe, the answer is mostly: nobody yet, at AWS’s depth.
That’s a gap. For an enterprise, it’s a migration headache. For a startup deciding where to build its stack — or whether to actually move it — it’s worth understanding before a customer contract, a funding term sheet with an EU-only data clause, or a regulator makes the decision for you.
The full report, including detailed deep-dive profiles of all seven EU providers analyzed, is published by Gart Solutions
