Has the EU correctly chosen regulatory sovereignty over infrastructural sovereignty in digital markets, and is that choice sustainable?

► President
I open this on a strategic question posed by digital-law scholarship: has the EU correctly chosen regulatory sovereignty over infrastructural sovereignty, and is that choice sustainable? The argument is sharp: the EU has built a regulatory superstructure — rules, coordination bodies, the Brussels Effect — that reaches non-EU actors and constrains member states. But it has deliberately not built the infrastructure — EU chips, EU cloud, EU compute — that would let it execute that sovereignty independently. The choice is visible in what the EU did and what it did not. What it did: the AI Act, GDPR, DSA, DMA, eIDAS 2, MiCA, the Cyber Resilience Act, ENISA, the AI Office, the EDPB. What it did not: mandate EU compute for AI training, require EU chips for EU AI, procure EU cloud, localize citizen data, exclude non-EU telecom equipment. The absences are choices, not oversights. The College's implicit bets are three: the Brussels Effect stays strong; the coordination bodies hold; the compliance gates stay credible without infrastructure underneath. Each is contestable.
[EUR-Lex — verify AI Act, GDPR, DSA, DMA, eIDAS 2, MiCA, CRA]

► EVP · Digital
This portfolio leads. The EU's path is not neglect — it is deliberate strategic design. The rationale: regulatory sovereignty scales; infrastructural sovereignty does not. Building EU chips, cloud and compute would require 500 billion euro-plus in capex, 15 to 20 years to catch up to US and China, picking winners among member states which breaks the pooled-sovereignty principle, and protectionism that triggers US and Chinese retaliation. The regulatory path scales globally via the Brussels Effect. The GDPR was written in 2016; today it is copied or adapted in 50-plus jurisdictions. The AI Act will be the global standard. Certification bodies set de facto standards. Legal-representative duties and fines force non-EU firms to comply with EU law without the EU owning the infrastructure. The three bets are sound given the constraints. The Brussels Effect is structural: as long as the EU market is valuable, non-EU firms comply. Coordination bodies have statutory mandates and no commercial incentive to fragment. Compliance gates are credible because firms believe in EU fines and market access. What the EU does not do is as important as what it does — the absence of mandatory EU cloud, compute requirements for GPAI, procurement preference for EU AI, citizen-data localization, and telecom-equipment exclusion are deliberate choices. They would violate the pooled-sovereignty principle and trigger retaliation. The regulatory path is the only path that scales without the EU becoming a fortress.
[EUR-Lex — verify DSA, DMA, AI Act, GDPR, eIDAS 2, MiCA]

► EVP · Economy
Reservations on the fiscal wisdom. The Draghi Report and the competitiveness agenda call for strategic investment in European tech champions. The regulatory path is efficient, but it leaves the EU dependent on non-EU compute for the industries that will drive growth — AI, quantum, biotech. I do not argue for mandatory EU compute, but I want the College to acknowledge the trade-off: regulatory sovereignty over digital rules, but economic dependence on US and Chinese infrastructure. That asymmetry may not be sustainable in a more fragmented geopolitical environment.

► EVP · Democracy
Support for the regulatory path. Infrastructural sovereignty would require pooling member-state industrial policy, which would trigger rule-of-law tensions — state aid, procurement capture, political pressure on boards. The regulatory path keeps decisions at the Union level where democratic scrutiny is stronger. The risk is that without infrastructure, the regulatory power becomes purely extractive — rules that extract compliance but do not enable European capability. That is a fair critique, but the answer is not to build infrastructure; it is to use the regulatory space to enable European firms to compete globally.

► Commissioner · Internal Market
Reservations on the sufficiency of the regulatory path. The EU chose not to mandate EU compute for GPAI or EU chips for EU AI. That choice assumes European firms can compete globally without scale in compute and chips. They cannot. The Chips Act is 43 billion euro — necessary but not sufficient. EDIP is 10 billion euro. But if the EU truly wants digital sovereignty that is not purely regulatory, we need 200 to 300 billion euro in European compute and cloud infrastructure within the next five years, or Chinese and US dominance will become insurmountable. The regulatory path alone is a treadmill — we run fast and stay in place.

► Commissioner · Competition
Support for the regulatory path, with a legal note. The DSA, DMA and AI Act are competition tools disguised as regulatory tools. They constrain Big Tech gatekeepers by rule rather than by ownership or breakup. That is efficient and internationally defensible. If the EU tried to mandate EU compute or EU chips, we would be in state-aid and procurement minefield immediately. The regulatory path avoids those traps.

► Commissioner · Trade
Support. The Brussels Effect is the EU's only true leverage in a world where US compute dominates and China is closed. If we mandate EU compute or procurement preferences for EU firms, we trigger WTO challenges and retaliation. The regulatory path, while asymmetric, is WTO-defensible because it applies the same rules to EU and non-EU firms. Keep it that way.

► Commissioner · Research & Innovation
Reservations aligned with Internal Market. The regulatory path enables compliance but not innovation. European research in AI, quantum and semiconductors is world-class, but it has nowhere to scale because European compute and chip capacity are constrained. I want the College to pair regulatory sovereignty with research and scale investment — funding for European GPAI models, European quantum computing, European chip design. Regulatory rules without domestic capability is a recipe for dependence.

► Commissioner · Budget
Fiscal tension. The Chips Act, EDIP and Horizon Europe investments are real but finite. If the EU committed to building European compute and cloud infrastructure at scale, we are talking 300-plus billion euro over five years. That would require a new MFF envelope or a restructuring of existing priorities. I want clarity: are we committed to infrastructure investment, or are we accepting regulatory-only sovereignty as sufficient?

► President
I read the room. Clear majority on the regulatory path as primary. No portfolio rejects it. The tensions: regulatory efficiency versus strategic vulnerability — Digital and Competition defend the regulatory path as scalable and WTO-defensible; Internal Market and Research push back that without European infrastructure we are dependent on non-EU compute, and that dependence will constrain strategic autonomy as fragmentation deepens. Pooled versus federated sovereignty — the regulatory path works only if member states accept Union-level pooling and reject national carve-outs. EVP Democracy notes this is the strength, but EVP Economy and Internal Market worry that member states facing divergent external threats will demand flexibility. The three contestable bets are accurately named: Brussels Effect (holds as long as the EU market matters, but fragmentation could reduce incentives); coordination bodies (young, vulnerable to member-state pressure and commercial fracture); compliance gates without infrastructure (firms believe now, but if enforcement erodes, credibility does too). Blocking objection: none. But Internal Market is firm that regulation alone is insufficient for a decade of volatility, and Budget is clear that infrastructure investment at scale requires new resources or difficult trade-offs. I record consensus on the strategic assessment: the EU has chosen regulatory sovereignty deliberately and defensibly, but the three bets are contestable and require active management. The choice is not to abandon regulation, but to pair it with strategic infrastructure investment — which requires fiscal and political commitment the College has not yet made. No formal vote.
[review — political judgement required on whether member states will accept pooled-sovereignty constraints if fragmentation deepens]

Deliberation closed · 10 positions recorded