The ideas here in this post are not without precedent. I look up to initiaves like Gaia-X with cautious optimism. However, I believe it is important, nevertheless, to put them out in the public and discuss them ever so often.
Like roads, ports, and power grids, data centers have become critical infrastructure. They underpin almost every function of a modern economy: commerce, communication, healthcare, public administration, defense, and scientific research. Treating them as optional or purely private assets no longer reflects economic reality. The question is therefore not whether they are essential, but whether it is prudent to rely almost exclusively on private, often foreign, providers to operate them.
There are two compelling reasons for the state to offer a public option in data center and cloud infrastructure (amongst others).
First, a public option would introduce natural competitive pressure into a market that is increasingly concentrated. Hyperscale cloud providers benefit from extreme economies of scale, network effects, and high switching costs, which together weaken meaningful price competition. This creates a real risk of price gouging, vendor lock-in, and unilateral changes to terms of service that users, especially small and medium-sized enterprises (SMEs) and public institutions, are powerless to resist. A state-backed alternative does not need to dominate the market to be effective; it only needs to exist as a credible option to discipline pricing and behavior across the sector.
Second, a public offering would provide a genuine guarantee of service for critical systems. Certain workloads (public registries, healthcare platforms, emergency services, scientific archives, and strategic industries) should not be subject to abrupt commercial pressures, geopolitical risk, or shareholder-driven priorities. By enabling these systems to be hosted on publicly owned infrastructure, states can ensure long-term continuity, transparency, and sovereignty, while still benefiting from modern tooling and professional operations.
It is increasingly popular, particularly in open-source circles, to imagine a future in which self-hosting replaces large-scale cloud services. While admirable in spirit, this vision underestimates the economic reality of infrastructure. Self-hosting loses competitiveness precisely where reliability, redundancy, security, and energy efficiency matter most. These are domains where economies of scale are decisive. Expecting individuals, nonprofits, or small organizations to replicate them independently is neither realistic nor efficient. The economies of scale are just not on the side of this strategy.
A state-backed option represents a pragmatic middle ground. Like public transportation, water utilities, or postal services, it leverages collective funding through taxation to achieve scale that no individual contributor could reasonably attain. Crucially, this does not preclude private innovation or competition. Instead, it ensures that essential services remain accessible, affordable, and more resilient.
Part of the resistance to this idea stems from the persistent mystification of data centers and web services. There is a widespread belief that only “Big Tech” can operate them competently. In reality, the technical knowledge required has never been more accessible. Decades of best practices, open standards, and free documentation are available to anyone willing to apply them. What Big Tech primarily offers is not secret knowledge, but capital concentration and scale (both of which governments already possess). I believe this is an important notion we need to spread more diligently. Without a doubt, new agencies created for these purposes will make mistakes, but they will be necessary learning steps towards the provision of an essential service.
At the European level, the opportunity is especially clear. Twenty-seven economies, aligned by regulatory frameworks and shared interests, could establish interoperable, publicly owned infrastructure following common standards. Such an initiative would reduce dependency on American providers, strengthen digital sovereignty, and dramatically improve access to high-quality networking and computing services for SMEs, startups, and public institutions.
However, such an effort would need to begin at the national level to prove its viability. Pilot projects, limited-scope public clouds, and targeted use cases would allow governments to validate the model before broader adoption. But the potential upside (economic resilience, strategic autonomy, fairer competition, long-term cost control) is substantial.
In short, treating data centers as public infrastructure is neither radical nor unprecedented. It is a rational response to their growing centrality in modern life. The question is no longer whether states can do this, but whether they can afford not to.