Progress once promised to open the world. It is now multiplying its locks. From digital infrastructure to strategic resources, the future is increasingly shaped by a decisive form of power: the ability to grant, restrict, or deny access.

For a long time, the future was imagined as a territory to be conquered. Each generation was expected to push beyond the boundaries inherited from the previous one, acquire greater knowledge, gain access to new possibilities, and leave its successors a world less constrained than the one it had known. Technological progress, education, industrialization, and the expansion of trade appeared to be part of the same historical movement: the gradual enlargement of human freedom.

This vision was never entirely consistent with reality. Major industrial transformations have always generated dependencies, monopolies, and forms of exclusion. Yet a significant difference is emerging today. Contemporary societies no longer depend merely on the resources required for their development. They increasingly depend on systems whose infrastructure, operating rules, and sometimes even decision-making processes remain beyond their control.

The future is no longer presented simply as a collection of possibilities. It is gradually taking the form of an architecture of permissions.

To communicate, work, produce, travel, access a service, or participate in the digital economy, individuals and organizations must now pass through a succession of technical, commercial, and institutional gateways. Each has its own conditions, intermediaries, and control mechanisms. Most remain invisible as long as everything functions properly. They suddenly become tangible when an account is suspended, a license expires, a provider interrupts a service, or a geopolitical restriction prevents access to a technology.

Power no longer resides exclusively in the ownership of the means of production. It also lies in the ability to determine who may use them, under what conditions, and for how long.

Ownership Gives Way to Access

One of the most profound transformations in the contemporary economy concerns the relationship between individuals and the goods they use.

For much of industrial history, acquiring an object meant obtaining a relatively autonomous capability. A machine could be repaired, a book preserved, a record passed on, and a vehicle maintained independently of its manufacturer. This autonomy was never absolute, but material ownership provided a degree of continuity between purchase and use.

The digital economy has progressively altered that relationship.

Software can be used without ever being owned. A library of films can disappear when a subscription ends. A vehicle feature may depend on commercial activation. A perfectly functional device can lose part of its usefulness when its manufacturer stops providing updates or disables a remote service.

Consumers no longer necessarily pay to possess a capability permanently. They pay to retain the right to exercise it.

This development has an economic rationale. Subscription models can finance continuous maintenance, distribute improvements rapidly, and make services accessible that would otherwise require prohibitive upfront expenditure. They can also reduce certain ownership costs.

But they introduce a new dependency: the continuity of use rests on a permanent contractual and technical relationship.

The object remains in the hands of its user, while part of its functionality remains in those of its provider.

This separation between possession and control is one of the defining characteristics of the contemporary economy. It gradually transforms ownership into revocable access, often without the full implications of that transformation being recognized.

The Invisible Infrastructure of Power

The global economy rests on an extensive network of infrastructure whose concentration has become a strategic issue.

Data centers, telecommunications networks, operating systems, payment infrastructure, distribution platforms, and computing services form the foundations of economic activity that appears, on the surface, increasingly decentralized.

An entrepreneur can establish a company from home, sell products across several countries, and use tools once reserved for large organizations. Yet this apparent freedom depends on a succession of intermediaries: hosting, identification, payments, distribution, advertising, storage, and data processing.

The number of visible participants can increase even as the essential infrastructure becomes more concentrated.

According to Synergy Research Group, Amazon, Microsoft, and Google together accounted for nearly two-thirds of the global cloud infrastructure services market in the first quarter of 2025. This concentration does not mean that alternative providers are absent or that every customer is trapped within a single system. It does, however, indicate that a substantial share of global computing capacity depends on a limited number of operators.

The problem is not simply their size. It concerns the cost of leaving.

When a company builds its applications around proprietary services, accumulates data in specific formats, and organizes its operations according to one provider's architecture, changing infrastructure becomes progressively more expensive. Migration costs, operational disruption, and the need to rebuild certain functions can turn an initially rational decision into a lasting dependency.

This phenomenon, known as vendor lock-in, does not necessarily result from abusive conduct. It can emerge from ordinary technical choices accumulated over several years.

Its consequences nevertheless extend beyond the commercial relationship.

A public administration unable to relocate its data, a hospital dependent on software that is difficult to replace, or a company incapable of operating without a remote service possesses only conditional autonomy.

The question of sovereignty consequently shifts. It no longer concerns only the territory on which infrastructure is located, but the practical ability to control, audit, and replace it.

Artificial Intelligence and the Concentration of Capabilities

Artificial intelligence intensifies this transformation because it shifts an increasing share of economic value toward resources that are difficult to reproduce.

The most advanced models require scientific expertise, datasets, electrical infrastructure, computing facilities, and substantial investment. Developing frontier systems therefore depends not only on the quality of an idea or the talent of a research team.

It also depends on access to the capabilities required to implement that idea.

The Stanford AI Index 2025 estimated that industry produced approximately 78% of notable artificial intelligence models in 2024, compared with 22% attributed to academia. This distribution reflects, among other factors, the growing importance of industrial resources in frontier research.

Several realities must nevertheless be distinguished. Open models, declining inference costs for certain tasks, and improvements in algorithmic efficiency are allowing new participants to access capabilities that were previously far more expensive. Artificial intelligence is therefore not moving exclusively toward closure.

But the democratization of use can coexist with the concentration of fundamental infrastructure.

A small company can integrate a powerful model into its products without possessing the resources necessary to train a comparable system. It can create value, innovate, and compete while remaining dependent on its provider's pricing, technical conditions, and contractual terms.

The distinction between the ability to use a technology and the ability to control it becomes essential.

A society can adopt artificial intelligence on a massive scale without developing an equivalent degree of technological autonomy. A state can modernize its public administration through foreign models while increasing its dependence on infrastructure beyond its jurisdiction.

Progress for the user does not necessarily translate into sovereignty for the institution.

As artificial intelligence enters manufacturing, finance, healthcare, education, and public administration, the question is no longer simply who possesses the most powerful models. It becomes who can modify the conditions of access, interrupt their use, or determine their limits.

The Geopolitics of the Keys

This new organization of power extends far beyond technology companies.

Governments have recognized that economic interdependencies can become instruments of foreign policy.

American restrictions on exports of certain advanced semiconductors and manufacturing equipment to China illustrate this development. They are designed to limit access to technological capabilities considered strategically sensitive, particularly for military applications and certain artificial intelligence systems.

China, for its part, possesses leverage associated with the processing of several critical raw materials, industrial supply chains, and its manufacturing weight.

Trade restrictions, investment screening, and technology localization policies reflect a common realization: in a highly integrated economy, controlling a component that is difficult to substitute can provide disproportionate influence.

Globalization was widely presented as a system in which expanding trade would reduce the likelihood of confrontation. It has indeed created interdependencies capable of supporting cooperation and making certain conflicts more costly.

But it has also produced indispensable points of passage.

International financial networks, advanced technologies, maritime routes, industrial equipment, and certain energy resources can become instruments of pressure when an actor possesses the ability to restrict access.

This does not mean that interdependence necessarily produces domination. Economic relationships can be reciprocal, and restrictions often impose costs on those who introduce them.

It means that dependencies are not always symmetrical.

A country possessing a technology that is difficult to replace does not occupy the same position as one that must import it to sustain industrial activity. A state with several suppliers retains more options than one tied to a single infrastructure provider.

Contemporary sovereignty is therefore measured less by declarations of independence than by the existence of credible alternatives.

When Identity Becomes Permission

The transformation also affects individuals.

Digital identity, authentication systems, biometric technologies, and automated verification procedures occupy an increasingly important place in access to services.

Their development responds to genuine needs. Fraud prevention, account protection, payment security, and administrative simplification require reliable identification mechanisms.

A properly designed digital identity system can facilitate access to rights, reduce administrative burdens, and enable previously excluded populations to participate in economic life.

But the generalization of these mechanisms raises a fundamental institutional question: what happens when identification becomes a prerequisite for almost every activity?

In a highly digitized environment, losing access to an account or encountering a verification error can have consequences far beyond the service concerned. It can prevent a payment, interrupt a professional relationship, complicate an administrative procedure, or temporarily make part of everyday life inaccessible.

The challenge is therefore not to reject digital identification. It is to ensure that the systems organizing it provide avenues of appeal, correction mechanisms, and appropriate alternatives.

A technical error should not become a form of social exclusion.

This requirement becomes even more important when automated decisions are involved in assessing applications, detecting suspicious behavior, or allocating certain services.

The efficiency of a system is not sufficient to establish its legitimacy. Its decisions must also be understandable, contestable, and subject to correction when they are wrong.

Without these safeguards, administrative simplification can produce a paradoxical form of bureaucracy: less visible, faster, but sometimes more difficult to challenge.

The Economics of Permanent Dependence

The concentration of infrastructure and the expansion of conditional access are changing the mechanisms of competition.

In a conventional economy, competition concerns prices, quality, availability, and product performance. In an economy organized around platforms and ecosystems, it also concerns the ability to retain users over time.

Companies naturally seek to establish stable commercial relationships. They invest in complementary services, integrated interfaces, and experiences designed to reduce friction.

These strategies can deliver genuine benefits. A coherent environment simplifies operations, improves compatibility, and reduces the need to manage multiple tools.

But when switching costs become excessive, loyalty ceases to be entirely voluntary.

Users may remain within an ecosystem not because they prefer it, but because their data, purchases, habits, and devices have become difficult to transfer elsewhere.

Dependence then becomes an economic asset.

It can strengthen pricing power, reduce competitive pressure, and discourage new entrants. In some markets, network effects amplify this process: the more users a service attracts, the more useful it becomes, and the more difficult it is for competitors to offer a credible alternative.

This dynamic is not uniform. Some markets remain contestable, open technologies can lower barriers, and rapid innovation can undermine positions that once appeared secure.

Economic history also demonstrates that technological monopolies are not necessarily permanent.

But it shows that challenging them may require major industrial, regulatory, or technological transformations.

Competition therefore depends not merely on the number of companies operating in a market. It depends on the practical ability of users to leave, new entrants to grow, and infrastructures to communicate with one another.

Economic freedom requires the ability to exit.

The Paradox of Abundance

There is an even deeper contradiction.

Never have human societies possessed so much accessible knowledge, so many communication capabilities, or such powerful production tools. An individual equipped with a computer can now perform tasks that once required an entire organization. Medicine advances, computing capabilities expand, and communication technologies overcome part of the constraints imposed by geography.

Yet this multiplication of possibilities does not automatically produce an equivalent expansion of autonomy.

A company can become more productive while depending increasingly on external providers. An employee can gain access to more powerful tools while exercising less control over the technical conditions of work. A country can become better connected to the world while growing more vulnerable to decisions made in other jurisdictions.

The paradox lies in the difference between power and control.

Power refers to what a system makes possible. Control refers to the ability to understand that system, govern it, and continue operating when it changes or disappears.

Contemporary societies are rapidly increasing their collective capabilities. They are not always developing the means to retain control over them at the same pace.

This divergence may represent one of the most important political transformations of the twenty-first century.

Technological progress expands the range of possible actions, while economic organization can concentrate decisions governing those actions.

Humanity gains capabilities without necessarily distributing the authority that accompanies them.

Autonomy Is Not Autarky

Faced with this development, the temptation is to demand complete independence.

Governments seek to relocate certain industries, companies diversify suppliers, and institutions develop their own infrastructure. These efforts respond to legitimate concerns, particularly when critical functions are involved.

But absolute autonomy is rarely a realistic objective.

Modern technologies depend on complex value chains, international specialization, and investments whose complete duplication would be extraordinarily expensive. Attempting to manufacture every component domestically, control every infrastructure layer, and eliminate every dependency could reduce economic efficiency without guaranteeing greater security.

The objective is therefore not to eliminate interdependence. It is to make it governable.

This requires several conditions: supplier diversification, system interoperability, effective data portability, contractual transparency, national or regional capabilities in critical sectors, and the ability to maintain essential services when disruptions occur.

Competition regulation also plays a central role. In Europe, the Digital Markets Act seeks, among other objectives, to limit certain practices by large platforms designated as gatekeepers. The Data Act aims to improve conditions for accessing and sharing data and for switching providers of certain data-processing services.

These instruments cannot, by themselves, guarantee the disappearance of dependencies. Their effectiveness depends on enforcement, market developments, and the capacity of economic actors to provide viable alternatives.

They nevertheless reflect an important conceptual shift: economic freedom is no longer reducible to the theoretical possibility of choosing. It requires that choices can actually be exercised under reasonable conditions.

A door is genuinely open only when it can be passed through.

The Democratic Question

Beyond economics and technology, the concentration of access mechanisms raises a democratic question.

Modern societies have built institutions around principles intended to constrain the exercise of power: separation of functions, accountability, judicial oversight, transparency of rules, and the possibility of challenge.

Yet an increasing share of the decisions shaping everyday life is made within contractual and technical environments whose mechanisms largely escape public debate.

Private companies must naturally be able to define the conditions of their services, protect their infrastructure, and enforce their legal obligations. Not every commercial decision is a matter of political sovereignty.

But when certain platforms become indispensable to communication, commerce, information, or professional activity, their decisions acquire collective significance beyond the ordinary relationship between a provider and a customer.

The difficulty lies in regulating this power without turning digital infrastructure into an instrument of state control.

Governments themselves can use identification, surveillance, and access-restriction technologies to limit the freedoms they claim to protect.

The problem, therefore, is not simply one of corporations versus governments.

It concerns the concentration of power, regardless of the institution exercising it.

A digital architecture genuinely compatible with civil liberties must allow private actors to be held accountable without creating new arbitrary powers for public authorities.

It must also preserve individual rights against both.

The fundamental question becomes one of accountability: who decides, according to which rules, under what oversight, and with what opportunities for appeal?

What It Still Means to Own the Future

Major periods of technological transformation have always generated concerns about the concentration of power.

Industrialization strengthened certain monopolies before encouraging new forms of competition and regulation. Electrification transformed relationships between public infrastructure and private enterprise. Telecommunications created networks whose collective importance gradually required particular obligations.

The present era is therefore not an absolute break with history.

Its distinctive feature lies rather in the extent to which systems have become integrated.

A single digital environment can now participate in communication, economic activity, identity, mobility, healthcare, and access to information. Dependencies are no longer isolated. They can reinforce one another.

This convergence makes the question of control more difficult to resolve.

The solution is not to choose between a completely open world and a completely secure one. No complex system can function without rules, safeguards, and authorization mechanisms.

The challenge is to determine which restrictions are necessary, which are proportionate, and which primarily serve to consolidate an existing position of power.

It also requires distinguishing freely accepted dependencies from those that become practically impossible to avoid.

Freedom does not require personally owning every infrastructure on which one depends. It requires not being entirely at the mercy of those who control them.

This distinction should guide industrial policies, digital strategies, competition rules, and the design of public services.

For the future will not be determined solely by the technologies societies succeed in inventing. It will also depend on the institutions they build to govern their use.

The question is not whether doors will continue to exist. Every complex civilization has them, and some are indispensable to its security.

The real question is who holds the keys, who can change the locks, and what happens to those left outside.

For a long time, progress was measured by the number of doors humanity managed to open.

From now on, it must also be measured by its ability to prevent them from closing.

Principal Sources

  • Stanford University, Institute for Human-Centered Artificial Intelligence, AI Index Report 2025. Data on notable AI model production, investment, performance, and costs.
  • Synergy Research Group, Q1 2025 Cloud Market Growth Stays Strong, 2025. Estimates of global market shares among leading cloud infrastructure providers.
  • European Commission, Digital Markets Act, Regulation (EU) 2022/1925. Regulatory framework governing large platforms designated as gatekeepers.
  • European Union, Data Act, Regulation (EU) 2023/2854. Provisions concerning data access, data use, and switching between data-processing service providers.
  • Bureau of Industry and Security, U.S. Department of Commerce, regulations and communications concerning export controls on advanced semiconductors and associated technologies destined for China, 2022–2025.
  • OECD, Digital Economy Outlook 2024, Volumes 1 and 2. Analysis of digital economic transformation, connectivity, and governance challenges.
  • World Bank, World Development Report 2021: Data for Better Lives. Analysis of data governance, power asymmetries, and the conditions for balanced economic and social use of data.

Methodological note: Quantitative figures refer to the editions and periods specified. The discussion of ownership, autonomy, concentrated power, and the future of individual freedoms constitutes an editorial analysis informed by these developments, rather than conclusions directly attributed to the organizations cited.