Matthias Bauer / Jul 2025
Image: Shutterstock
Europe’s ambitions in digital sovereignty, AI, and innovation are growing louder by the day – but much of it remains vague and inconsistent. While political rhetoric is abundant, clear definitions and coherent policies are often lacking. One core ingredient, however, is still missing from this recipe for success: real cloud customer choice.
Our new paper assesses three fundamental policy domains – regulation, competition enforcement, and standard-setting – and how each can either advance or inhibit customer choice in the cloud market. Unlike many policy debates that fixate on how rules can be weaponised against specific firms, our analysis focuses on something far more constructive: how to make the entire cloud ecosystem work better for users.
Why does cloud customer choice matter? Because it unlocks everything else. The ability to freely select, combine, and switch between cloud services is essential for building the complex, interoperable data infrastructures that AI and digital transformation now require. Without it, users – from SMEs to public authorities – face vendor lock-in, high switching costs, and shrinking negotiating power. And Europe risks remaining stuck in a cycle of siloed, shallow digital adoption.
Let’s be clear: cloud customer choice is not about picking national champions or excluding foreign providers. It is about maximising value, flexibility, and performance for users – regardless of origin. Real sovereignty lies in the freedom to choose what works best. That principle must now guide Europe’s approach to cloud governance – and, more broadly, to technology adoption as a whole.
Yet today, three key obstacles stand in the way:
- Fragmented and overprescriptive regulation
The EU’s own rules – including the Data Act, the Digital Markets Act (DMA), and emerging proposals like the Digital Networks Act (DNA) – often conflate good intentions with clumsy execution. In trying to enforce switching rights and interoperability, the Data Act ends up imposing rigid unbundling mandates that ignore technical realities. Originally designed to regulate two-sided digital platforms, the DMA extends gatekeeper obligations to enterprise cloud services by classifying them as core platform services – a mismatch that has not yet been tested through enforcement, but is already generating significant legal uncertainty. The forthcoming DNA risks regulating cloud services like telecom networks – a category error that could stifle investment and innovation.
Instead of reinforcing legal confusion, EU regulation should support user-centric outcomes: portability, contract flexibility, and service interoperability. Rules must reflect the actual diversity of cloud services – not reduce them to lowest-common-denominator regulatory templates, as seen in large parts of the DMA or the Data Act.
- Case-by-case competition enforcement
While national competition authorities have grown more active – rightly investigating bundling, licensing, and market foreclosure – their approach remains reactive and legally constrained. They cannot redesign cloud market structures from the top down. Nor should they.
What they can do is provide clearer guidance on how modern market practices – such as bundling, architectural integration, or pricing structures – relate to user switching and market contestability. This includes recognising when API alignment or vertical integration is pro-competitive, especially where it improves interoperability.
Equally important is avoiding overlap between competition law and new regulatory regimes like the DMA – a risk that could lead to double jeopardy for companies and encourage legal forum shopping by authorities.
- Lack of global, open, non-discriminatory technical standards
Standardisation offers perhaps the greatest untapped opportunity to improve customer choice – but only if it is done right.
We argue for a more active embrace of market-led and open-source-compatible standard-setting, including the development of FRAND-style licensing principles for cloud and AI services. This would provide clarity on acceptable contractual terms, support SME participation in standard-setting, and reduce legal and commercial uncertainty.
Crucially, standard-setting must not be hijacked by closed national initiatives or bureaucratic platforms that fail to deliver global adoption (think: GAIA-X). Nor should interoperability be imposed top-down through rigid regulation that, once in place, tends to become entrenched – regardless of its fitness for evolving technologies.
Meaningful standardisation supports diversity – it does not erase it.
From rhetoric to results
The economic potential at stake is enormous. If Member States converge toward the current frontier in cloud and AI adoption, the EU could unlock up to €1.2 trillion in GDP gains by around 2030 – with €450 billion available to EU governments annually from public sector modernisation alone. But these gains require not just more data centres or new regulatory instruments – they require an ecosystem built on flexibility, competition, and the objective of maximising user choice.
That is why we call for a dual-track strategy: one that combines near-term policy wins with long-term institutional reform.
In the short run, this means clearer enforcement guidelines, procurement incentives for multi-cloud compatibility, and better support for open standardisation efforts.
Over the longer term, it means redefining “sovereignty” away from nationality and towards freedom of contract and infrastructure portability.
The future of Europe’s cloud and AI economy will not be built through slogans or protectionist shortcuts. It will be built through deliberate, technically grounded choices that empower users. Customer choice is not a side issue – it is the cornerstone of competitiveness.
