Who determines
the conditions?

A three-week sequence

On 3 June 2026, the European Commission presented its Technological Sovereignty package: it proposes to turn the sovereignty requirement, until now a matter of voluntary certification, into a binding condition of access to public procurement of cloud computing services, graduated according to the sensitivity of the uses. On 24 June, the same Commission presented a new simplification package comprising two legislative proposals, pursuing a programme launched in February 2025, numbering a dozen proposals to date, and whose first substantial piece, definitively adopted in February 2026, undoes vigilance obligations the Union had adopted in 2024.

Within three weeks, the same institution proposes to attach conditions with one hand and pursues their withdrawal with the other.

This back-and-forth is not an inconsistency of calendar. It makes legible a question the European debate does not ask: what act does a power accomplish when it attaches a condition, and what act does it accomplish when it withdraws one?

The grammar of capacities

The solutions proposed for Europe today resemble one another in their grammar. Channelling savings towards technology companies, subsidising manufacturing capacities, taking public stakes, building computing infrastructure: these measures differ in their instruments and converge in their operation. All increase a capacity within an existing order. More capital, more factories, more compute, measured against indicators that others have established.

This grammar is not peculiar to Europe. In Shanghai as in Washington, technological competition is now narrated in capacities: amounts committed, data centres, infrastructure shares, semiconductors. The question is almost always quantitative. The debate bears on amounts, vehicles, sectoral priorities. It never bears on the nature of the act.

The other act

There exists, however, a second type of act, irreducible to the first. Determining the conditions of others' action: fixing what an actor will have to comply with in order to access a market, sell a product, process data, qualify an investment. Increasing a capacity extends what Roman law called potestas, the power to execute within a given order; determining a condition touches auctoritas, the authority that carries the order itself. No addition of capacities produces, by itself, the power to condition.

The point this Position holds is then the following: the instruments are heterogeneous, but the act can be homogeneous. A technical standard, a sustainable finance taxonomy, a public procurement criterion, a portability obligation, a condition of access to the internal market belong to distinct policies, carried by different directorates, discussed in separate arenas. These instruments differ in their source, their regime and their immediate object. They nonetheless have a common legal effect: each creates an external condition that other actors must satisfy in order to pursue their own activity. It is this common effect, and it alone, that justifies reading them as a single category of acts. It is not a community of purpose; it is a community of nature. Recognising it changes the level of the discussion: it ceases to be a trade-off between public policies and becomes a question of qualification.

The precedent and what its withdrawal extinguishes

Europe is, of the three great powers, the one for which this question decides the most. The United States binds capital through ownership: the public stake at the table. China binds through control. Europe built its position on a third mode: the condition. Ownership and control survive the extinction of this mode; a position built on it does not.

This mode has a verifiable history. For years, companies established outside Europe adapted their products, their contracts and their processes in order to keep accessing the European market. What the literature has named the Brussels effect did not proceed from an additional capacity: standards designed for one market were integrated far beyond it, not because Europe was the most capacitated, but because it determined conditions that others had an interest in satisfying. This precedent establishes one thing: the power to condition is not proportional to capacity. It is of another nature, and it was, for a time, the European mode of exercising power.

It is in the light of this distinction that the June 2026 sequence becomes significant. That a simplification programme lightens declarative burdens is a public policy choice, debatable as such. But when simplification withdraws conditions to which global actors had conformed, it does not merely lighten a burden: it extinguishes an act. The question is not whether each withdrawn obligation was well calibrated. It is whether the institution that withdraws qualifies the nature of what it withdraws. Nothing, in the documents accompanying these reforms, indicates that the difference between reducing a cost and extinguishing a condition was formulated, still less qualified. The revision of the duty of vigilance, first substantial piece of the programme, was moreover prepared without an impact assessment or public consultation; the European Ombudsman found there, in November 2025, procedural failings constituting maladministration. The debate was held entirely in the register of capacity: burden, competitiveness, amounts saved. The object, for its part, touched the register of the condition.

There lies the diagnosis, and it carries no reproach. The economics of industrial policy has accomplished a remarkable empirical turn; it now measures finely the effect of instruments. But measuring and qualifying are not two competing answers to a single question: they are two levels of analysis. Measurement establishes that a condition produces effects; it presupposes that the condition exists. By what act a condition exists, what makes it executable, what distinguishes its withdrawal from a mere saving of costs: this question is prior to any measurement, and it does not belong to economics. It is not a refinement; it is the prerequisite. One can neither build nor dismantle lucidly a power of condition without having qualified the act that constitutes it. Europe, whose position rests precisely on this mode of exercise, is the power that can least dispense with this qualification, and the one that, in 2026, dispenses with it most visibly.

Closing

A power can increase all its capacities and determine the conditions of nothing. Sovereignty is not received: it is produced. It is produced in the act that conditions, and that act, today, remains without qualification in the debate that decides its fate.

The instruments are heterogeneous. The act, for its part, is homogeneous: determining the conditions of others' action.