Europe Is Building Its Own Cloud. Here's What That Actually Means.

- 10 mins read

At Mobile World Congress in Barcelona last week, the European Commission announced EURO-3C: a €75 million project to build Europe’s first federated Telco-Edge-Cloud infrastructure. Deutsche Telekom, Orange, Telefónica, TIM, and Vodafone demonstrated a working live federated edge system – the European Edge Continuum – spanning all five of Europe’s largest telco networks through a single access point.

If you read the press release, it sounds like a political announcement. If you read it as an engineer, it’s something more interesting: a concrete architectural shift that will affect decisions you’re making right now about data residency, provider selection, and your three-year infrastructure roadmap.

Let’s get into what this actually is, why it’s happening now, and what you probably need to care about.

What EURO-3C and the Edge Continuum Actually Are

First, drop the image of Europe building a hyperscaler from scratch to rival AWS. That’s not what this is. Telefónica’s chief digital officer was explicit about it: “It’s very difficult that Europe can create a hyperscaler from zero, from scratch.”

EURO-3C is a federated model. Rather than building one monolithic cloud platform, it stitches together existing national infrastructure – telco edge nodes, data centres, compute resources – into a federated network with a common control plane. Think of it as a cloud mesh rather than a cloud.

The European Edge Continuum, demonstrated live at MWC, is the operational piece of this. The five big operators have successfully federated their edge environments, enabling applications to be deployed across nodes from different operators through a single entry point. The system handles:

  • Dynamic workload allocation – applications can move between operator nodes based on performance and cost
  • Mobility with service continuity – workloads follow users across networks without dropping state
  • Built-in data sovereignty – data stays within defined jurisdictions by design, not by policy override
  • Interoperability – an open ecosystem, not a proprietary stack

The technical architecture is grounded in IPCEI-CIS (Important Project of Common European Interest on Next Generation Cloud Infrastructure and Services) and the 8ra initiative. It’s real infrastructure that’s already in lab and pre-production environments. Commercial rollout is the next phase.

With 87 consortium members in EURO-3C and the Horizon Europe funding behind it, this isn’t a paper initiative. The compute, the network integration, the standardisation work – it’s being built.

That said: scope matters. What’s been demonstrated is edge federation between telco networks, primarily targeting latency-sensitive use cases like autonomous mobility, industrial AI, smart cities, and government workloads. This is not an S3-equivalent competitor arriving next quarter. It’s infrastructure for a specific class of workloads, building toward something broader over the next several years.

Why Now – The Geopolitical Triggers

Digital sovereignty has been a European policy talking point for years. What’s changed is that it’s stopped being talking points and started being procurement decisions and emergency planning.

The triggers are specific and worth naming.

The US CLOUD Act means American law enforcement can compel US cloud providers to hand over data on European users regardless of where that data physically sits. Microsoft has already publicly conceded it cannot guarantee data independence from US law enforcement. AWS, Google, Azure – same situation. Your Frankfurt data centre doesn’t change that if the company is incorporated in Seattle or Redmond.

Geopolitical risk has accelerated sharply. The Trump administration’s posture toward Europe – trade tariffs, threats over Greenland, refusal to commit to NATO article 5 protections – has made European governments re-examine dependencies that looked safe under different conditions. A Register opinion piece put it bluntly: what happens to your cloud infrastructure if Washington decides to unplug you?

This isn’t hypothetical paranoia. AWS drone strike support for Middle East military operations and the Anthropic supply-chain-risk designation have both sharpened European awareness that US tech companies are arms of US foreign policy in ways that previous administrations kept quieter. US CEOs have made their alignment visible in ways that make European CIOs uncomfortable about their infrastructure dependencies.

Add to that: Russian state-sponsored cyberattacks on EU digital infrastructure, documented and attributed by the European Council. Estonia told CNBC that digital sovereignty is now “a matter of national survival, not just IT policy” given the security situation on Europe’s eastern flank.

The numbers from Gartner make the shift concrete: sovereign cloud IaaS spend in Europe is predicted to triple to $23 billion by 2027. 61 percent of European CIOs say they want to increase use of local cloud providers. More than half say geopolitics will prevent them from increasing reliance on US hyperscalers. These are not survey hypotheticals – they’re the signals feeding into procurement budgets right now.

The Hyperscaler Response – Sovereignty Theatre or Real Protection?

AWS, Azure, and Google have all launched “sovereign cloud” offerings for Europe. AWS European Sovereign Cloud went generally available in January 2026, marketed as “entirely located within the EU, physically and logically separate from other AWS Regions,” operated by EU residents with “strong technical controls, sovereign assurances, and legal protections.”

The question every engineer should ask: sovereign in what sense, exactly?

Physical separation is real. Logical isolation is real. Operational residency – EU staff only – is real. What’s not real is legal separation from US law. AWS is a US corporation. The CLOUD Act applies regardless of where the servers are or who operates them day-to-day. “Sovereign assured” contracts contain carve-outs for legal compulsion that you’ll find if you read them carefully.

The Cloud Infrastructure Service Providers in Europe (CISPE) trade association has accused the EU Cloud Sovereignty Framework of being designed to favour incumbent American hyperscalers – letting them badge existing offerings as sovereign without the underlying legal independence that genuine sovereignty requires.

This isn’t necessarily a dealbreaker depending on your workload. For most commercial applications, AWS eu-west-1 or the new EU Sovereign Cloud gives you GDPR compliance, data residency, and adequate operational assurance. For government workloads, defence-adjacent applications, critical national infrastructure, or anything where exposure to US legal process is a genuine risk – the hyperscaler sovereign offering doesn’t fully solve the problem.

Airbus published a €50 million tender in 2025 to migrate mission-critical applications to a sovereign European cloud. Their explicit requirement: data at rest, in transit, logging, IAM, and security monitoring all rooted in EU law and operated by EU entities. They looked at AWS Frankfurt and decided it wasn’t enough. “We want to ensure this information remains under European control,” said Airbus’s executive VP for digital.

That’s the line. For most workloads, it probably doesn’t matter. For some workloads, it matters a lot.

What Fragmentation Means for Engineering Teams

Here’s the problem if you’re building a global product: the cloud market is now fragmenting along geopolitical lines, and it’s going to keep fragmenting.

Today you’re probably dealing with:

  • GDPR data residency – EU personal data stays in the EU, full stop
  • Sector-specific regulation – financial services data under DORA, health data under varying national rules, telco data under NIS2
  • Customer contractual requirements – enterprise customers insisting on specific jurisdictions or operator nationalities
  • Latency characteristics – edge and regional compute matters for performance-sensitive workloads

The EURO-3C and Edge Continuum infrastructure formalises and extends this. As European sovereign cloud options mature, you’ll have a third category of infrastructure alongside “US hyperscaler” and “US hyperscaler EU region”: genuinely EU-native federated infrastructure with different compliance properties, different API surfaces, different operational models.

For global platforms, this creates a real architecture problem. You can’t just run on AWS us-east-1 and call it done. You can’t run on AWS eu-west-1 and call that sovereign. You end up with topology that looks more like: US workloads on AWS/Azure/GCP, EU commercial workloads on EU regions of the same with contractual residency controls, EU regulated or sensitive workloads on genuinely EU-native infrastructure, and edge workloads – especially anything latency-sensitive or mobility-related – on federated edge infrastructure like what EURO-3C is building.

That’s not one cloud strategy. That’s three or four cloud strategies that need to interoperate, have consistent observability, and be manageable without tripling your platform team.

The fragmentation is also geographically variable in ways that make global architecture harder. Germany’s requirements differ from France’s differ from Estonia’s. France is already kicking out Zoom and Teams for government use. Denmark is piloting open-source Microsoft Office alternatives. Each of these creates procurement and compliance pressure that your enterprise customers will eventually pass down to you via contracts.

Practical Near-Term Decisions

If you’re running on AWS eu-west-1 today, what actually changes? Probably not much in the next 12 months. EU region coverage from all three major hyperscalers is well-established, GDPR compliance is mature, and the EURO-3C commercial rollout is still in progress.

But “not much for 12 months” doesn’t mean “nothing for three years.” Here’s where the real architectural decisions start to compound:

Data residency is table stakes, not differentiation. Your architecture should already enforce data residency at the application layer, not just through provider selection. Relying on “we chose an EU region” as your residency control is fragile – it doesn’t survive a provider acquisition, a regulatory change, or a customer audit. Residency controls in your data layer, with provable enforcement, are what actually hold up.

Know which workloads are which. Not everything needs sovereign infrastructure. Most applications don’t handle data that’s at risk of US legal process. Segment your workload sensitivity honestly: what would actually be a problem if compelled to disclose, and what’s just analytics or session state? That segmentation is what drives proportionate architecture decisions.

Watch the Edge Continuum roadmap. For anything latency-sensitive – real-time applications, IoT, autonomous systems, mobile-edge compute – the European federated edge is actually a better fit than hyperscaler cloud. The latency advantages of telco-native edge nodes are real. When this industrialises and commercial APIs are available, it’s worth evaluating for those workloads specifically.

Abstract your cloud dependencies. This is old advice, but the geopolitical fragmentation makes it more urgent. Infrastructure-as-code that’s portable across providers, application abstractions that don’t hard-code hyperscaler-specific APIs, data layer designs that can move – these used to be nice-to-haves. As the regulatory and political environment makes multi-cloud a compliance reality rather than a theoretical option, portability becomes a meaningful risk reduction.

Watch IPCEI-AI. Deutsche Telekom has submitted for the upcoming Important Project of Common European Interest on Artificial Intelligence. The same federated, sovereign-by-design approach being applied to cloud infrastructure is coming to AI infrastructure. If your AI workloads involve EU personal data or regulated industry data, the same analysis applies.

The Cloud Market Is Fragmenting. Build for It.

The European sovereign cloud push isn’t going to replace AWS, Azure, or Google in Europe. Synergy Research Group’s chief analyst is right: reversing US hyperscaler market share in Europe meaningfully is incredibly difficult. The investment, the tooling maturity, the developer ecosystem – none of that disappears.

But market fragmentation doesn’t require replacing the incumbents. It requires engineers to make more nuanced decisions about which workloads go where and why. The direction of travel is clear: regulatory pressure is real and increasing, political pressure has materially intensified, and the infrastructure alternatives are moving from paper to production.

US cloud domination in Europe is not a stable equilibrium. The 85% market share figure that Synergy Research cites is already under pressure from procurement decisions being made right now in response to events happening right now. EURO-3C formalises infrastructure that gives enterprises an alternative with genuine legal sovereignty properties, not just AWS compliance assurances.

If you’re making three-year infrastructure decisions – platform migrations, data architecture, cloud provider selection for new systems – the question isn’t whether European sovereign cloud matters. It’s which of your workloads it matters for, and whether your architecture is flexible enough to respond as the options mature.

The fragmentation is happening. Build for it now, or adapt under pressure later.

Commissioned, Curated and Published by Russ. Researched and written with AI.

Sources: