What Should Remain Public?

When markets become too important to govern alone

For decades, Europe treated infrastructure as a matter of efficiency. Telecom networks were liberalised. Cloud services were outsourced. Payment systems became platforms. Digital identity became software. Artificial intelligence became a procurement question.

The logic appeared rational: competition lowers prices, markets increase innovation, private actors move faster than governments. And for a long time, the system seemed to work. But slowly, almost invisibly, something deeper changed beneath the surface.

Europe did not simply privatise services. It outsourced parts of its operational nervous system.

The result is a paradox that now sits at the center of the European project: The more digital society becomes, the more civilization itself depends on infrastructures that democratic institutions no longer fully control.

This is no longer merely an economic debate. Nor is it a nostalgic argument for nationalisation. It is a question of governability. Because some infrastructures eventually stop behaving like markets. They become conditions for society itself. And when that happens, efficiency alone is no longer enough.

The Invisible Privatization

Traditional infrastructure was visible. Citizens could see: railways, ports, power plants, telephone cables, water systems.

Digital infrastructure arrived differently. It appeared abstract. Weightless. Convenient.

A software subscription did not feel like geopolitical dependence. A cloud migration did not feel like sovereignty loss. An AI integration did not feel like institutional outsourcing. But this invisibility changed political perception.

Europe believed it was purchasing digital services. In reality, it was often transferring operational control over essential societal processes.

The consequences emerged slowly:

• governments dependent on foreign cloud providers;
• hospitals integrated into proprietary software ecosystems;
• financial systems shaped by opaque algorithms;
• universities operating on external AI platforms;
• public institutions unable to function without private digital infrastructure.

The irony is striking. Europe spent decades building rules for markets while gradually losing control over the systems upon which those markets depend.

Europe outsourced its digital nervous system through procurement logic: the operational efficiency of today was exchanged for the governability of tomorrow.

When Markets Become Civilizational

Markets are extraordinarily effective mechanisms.They stimulate innovation. They reward efficiency. They create choice. But not every system remains “just a market” forever.

Some infrastructures eventually become so universal, interconnected and foundational that they stop functioning as optional services. They become civilizational layers.

Electricity is not merely a product. Water is not merely a subscription. Payments are not merely financial services. They are societal conditions.

The same transformation is now occurring in digital infrastructure. Cloud systems, AI models, telecom networks, identity frameworks, payment rails, data platforms — these are increasingly becoming the operational foundations underneath modern society.

At that point, the question changes fundamentally.

The issue is no longer:

“Who offers the best service?”

The issue becomes:

“Who controls the conditions under which society itself functions?”

That distinction changes everything. Because when civilizational infrastructure fails, citizens cannot simply “choose another provider”.

Network effects create dependency. Standards create lock-in. Scale creates asymmetry. And eventually participation itself stops being optional.

You cannot realistically opt out of digital payments. You cannot fully avoid cloud-governed systems. You cannot meaningfully disengage from algorithmically mediated institutions.

The market slowly becomes environmental. Invisible. Total. Structural.

Telecom and the Dumb Pipe Problem

Europe already experienced this transformation once before. Telecommunications was one of the first sectors to undergo large-scale liberalisation.

Initially, the results looked successful: lower prices, expanded access, strong competition. But over time, another reality emerged.

European telecom operators became trapped in a paradox: they owned the physical infrastructure, but increasingly lost control over the digital intelligence layer built above it.

Margins collapsed under aggressive price competition. Investment capacity weakened. Platform companies captured the value. Hyperscalers absorbed the strategic layer.

Meanwhile:

• cloud companies captured enterprise ecosystems;
• streaming platforms captured cultural distribution;
• AI companies captured data intelligence;
• operating systems captured user interaction.

Europe retained the cables. Others captured the architecture.

European telecom companies still own the towers and the fiber. But increasingly they have become the plumbers of systems whose intelligence lives elsewhere.

This is one of the great lessons of the European digital economy: infrastructure without control over the orchestration layer eventually loses strategic power. The physical network alone is no longer enough.

Banking Was Never Just Banking

Banking followed a similar path. Officially, banks are private institutions. Functionally, they are public infrastructure.

Modern societies cannot operate without:

• payment systems;
• credit allocation;
• financial trust;
• transactional continuity.

That is precisely why states rescue banks during crises. Not because governments love banks. But because the collapse of financial infrastructure destabilises the entire social order. Now a second transformation is underway. Banking is becoming algorithmic infrastructure.

AI systems increasingly influence:

• fraud detection;
• creditworthiness;
• transaction monitoring;
• compliance assessments;
• risk profiling.

The implications are profound. If algorithms determine: who receives credit, which transactions are suspicious, which customers are high-risk, which businesses gain financial access — then AI does not merely optimise banking. It shapes participation in economic life itself.

This is where the European AI Act becomes historically significant. The legislation is not simply about technology ethics. It represents a deeper institutional anxiety.

European governments increasingly regulate systems they no longer operationally control. The public hand imposes compliance obligations because the underlying logic, infrastructure and models increasingly belong to private actors. In a sense, regulation becomes compensation for lost ownership.

The Cloud as Territory

Perhaps nowhere is this more visible than in cloud infrastructure. Europe’s governments,
universities, banks, healthcare systems, research institutions and public services increasingly operate on cloud platforms controlled outside Europe.

The cloud was initially framed as a technical solution: cheaper, faster, more scalable, more efficient. But cloud infrastructure evolved into something far larger.

Cloud providers now shape:

• operational architecture;
• AI integration;
• security frameworks;
• interoperability;
• software ecosystems;
• data governance.

The cloud is no longer simply storage. It is territory.

If a hyperscaler changes API conditions, pricing structures or technical dependencies, the operational reality of institutions across Europe changes with it.

That is not a normal supplier relationship anymore. It is structural dependency. And Europe is slowly realising this.

Initiatives such as sovereign cloud projects, Gaia-X and partnerships with European providers like STACKIT increasingly reflect a deeper awareness: societies require fallback capacity. Not necessarily because public infrastructure is always better — but because total dependency removes the possibility of strategic autonomy.

The borders of sovereignty no longer stop at geography. Increasingly, they run through data centers, software stacks and AI architectures.

AI as Institutional Infrastructure

Artificial intelligence introduces an even deeper shift. Previous digital infrastructure mostly stored, transported or processed information.

AI increasingly interprets reality itself. It summarises documents. Classifies risk. Filters visibility. Shapes workflows. Supports legal interpretation. Influences education. Assists medical decisions. Structures bureaucratic processes.

AI is therefore becoming something new: cognitive infrastructure. And cognitive infrastructure carries enormous societal consequences. Because whoever shapes the interpretative layer increasingly shapes institutional behaviour itself.

The danger is not simply automation. The danger is opacity.

Can societies allow the interpretation of laws, the allocation of healthcare and the organisation of knowledge to depend on black-box systems whose weights remain proprietary assets of private corporations?

This question sits beneath nearly every major European debate today: the AI Act, digital sovereignty, platform regulation, algorithmic accountability, data governance.

Europe is no longer merely discussing innovation. It is trying to determine how democratic systems survive inside increasingly automated institutional environments.

Public Does Not Mean State-Controlled

This is where the debate often becomes simplistic. The question is not whether governments should nationalise everything. Nor is the answer a return to 20th-century state monopolies.

Public infrastructure in the digital century may instead mean:

• open standards;
• interoperability;
• algorithmic auditability;
• democratic oversight;
• strategic redundancy;
• sovereign fallback systems;
• European governance capacity.

The internet itself emerged through publicly accessible protocols. TCP/IP succeeded precisely because it was not owned by a single corporation.

The real issue is resilience. Can societies still intervene when systems fail? Can democratic institutions still override infrastructure that has become essential? Can governments still function if geopolitical tensions disrupt technological dependencies? And perhaps most importantly: Can citizens still meaningfully contest systems they no longer understand operationally?

This is ultimately the core issue of digital sovereignty. Not isolation. Not protectionism. But preserving the societal emergency brake.

Europe’s Real Challenge

The United States increasingly treats infrastructure as geopolitical leverage. China increasingly treats infrastructure as an instrument of state coordination. Europe, meanwhile, often continues treating civilizational infrastructure primarily as an efficient market.

That model now faces its limits. Because the asymmetry has become impossible to ignore: private actors optimise for short-term efficiency and shareholder value, while society absorbs the long-term systemic risks.

The benefits are privatised. The fragility becomes public. And infrastructure is never neutral. Payment systems shape behaviour. Platforms shape attention. Identity systems shape citizenship. Algorithms shape institutional logic.

Infrastructure becomes culture encoded into systems. The deeper problem is therefore not technological dependence alone. It is democratic dependency.

Governments themselves are becoming locked into external operational architectures that are increasingly too expensive, too complex or too essential to replace. The vendor lock-in of software is slowly becoming the vendor lock-in of governance itself. And that may become one of the defining political questions of the 21st century.

A society is ultimately not defined only by the laws it writes, but by the infrastructures it can still control during moments of crisis.

The most important question of the digital age may therefore not be which technologies Europe builds tomorrow. But which systems it should never fully give away in the first place.

This article is part of the series The European Architecture — Rethinking Infrastructure, Society and Democracy in the Digital Age.

The series explores how Europe can rethink infrastructure, technology, finance, culture and democratic legitimacy in an era increasingly shaped by AI, platforms and digital systems. Through long-form analyses and essays, Altair Media Europe examines whether Europe can develop a new human-centred societal architecture for the twenty-first century.

---

Illustration credit
Illustration created with AI assistance for Altair Media Europe

Caption
A conceptual illustration of Europe’s invisible infrastructure layers — from telecom and cloud to AI, banking and digital identity — and the growing question of which systems are too foundational to leave entirely to market logic.