Ireland took over the rotating Presidency of the Council of the EU on July 1, 2026, inheriting a six-month chair of Europe's most consequential digital-infrastructure file: the Cloud and AI Development Act (CADA). The Commission published its formal proposal on June 3, 2026, as the centerpiece of the bloc's Tech Sovereignty Package. Dublin's own published Presidency programme is candid about how far it intends to push the file — officials plan to open "main discussions" among member states but aim only to submit a progress report by December, not a final Council position (Irish Presidency programme, via EU Tech Loop). That caution is not bureaucratic timidity. It reflects a structural conflict of interest that no amount of procedural neutrality can fully paper over.
What CADA Actually Does
CADA is not a paperwork exercise. The Commission's own framing sets a target to triple the EU's data-centre capacity within five to seven years, paired with a four-tier cloud "assurance" scheme that public-sector buyers must use when procuring cloud services (European Commission, Digital Strategy). Under the framework, a US hyperscaler partnered with an EU company could qualify for Level 2; Level 3 and 4 status — reserved for the most sensitive public-order workloads — effectively requires EU ownership, EU-based personnel, and control over the software supply chain (Legislative Train Schedule, European Parliament). That is a real, if narrow, procurement lever aimed squarely at reducing European public-sector dependence on Amazon Web Services, Microsoft Azure and Google Cloud.
The case for it deserves a fair hearing. Europe genuinely lacks a hyperscale cloud champion of its own, and relying on a handful of US firms for public-sector and defense-adjacent workloads creates a real single point of failure — one sharpened by the extraterritorial reach of US law over data held by US-headquartered companies, wherever the servers physically sit. A sovereignty tier for the most sensitive government workloads is a proportionate response to that risk, not paranoia.
The Country in the Chair Is the Country Most Exposed
But the country chairing this discussion is also the one with the most to lose from a heavy-handed version of it. Ireland hosts the European operations of sixteen of the world's twenty largest technology companies, including the European infrastructure backbones of AWS, Microsoft and Google, per Ireland's own Presidency materials (EU Tech Loop, citing the Irish Presidency programme). Corporation tax from a small number of US multinationals funds a disproportionate share of the Irish exchequer. A Council chair pushing CADA's sovereignty tiers too aggressively would be legislating against its own tax base — which is precisely why Dublin has scoped its Presidency ambition down to a progress report rather than a deal.
This is not a new pattern. Ireland's Data Protection Commission has spent a decade as the EU's lead GDPR regulator for Big Tech precisely because those firms are headquartered there, and it has faced years of criticism — from privacy advocates and fellow regulators alike — that economic dependency has produced slower, softer enforcement than the law intended. Tech Policy Press has now made the same argument about the CADA Presidency: Ireland is "torn between a deep indebtedness to the European project and economic dependency on US multinationals," acting as honest broker on a file that touches the companies it cannot afford to alienate (Tech Policy Press).
The Capacity Constraint Nobody Can Legislate Away
There is a second, more immediate problem: Ireland could not host a tripling of data-centre capacity even if it wanted to. Data centres already consumed roughly 23% of Ireland's national electricity in 2025, growing from just 5% a decade earlier and accounting for 85% of the country's total demand growth over that period, according to the energy regulator's own figures (CRU, Large Energy Users Connection Policy decision, 12 December 2025). EirGrid, the state grid operator, has told developers not to expect new data-centre connection approvals in the greater Dublin area until 2028 (Data Center Dynamics). The CRU's response was not a ban but a conditional reopening: new large energy users must now bring their own new renewable and dispatchable generation to the grid rather than simply drawing on it — a sensible, market-based fix to a real physical constraint, not an ideological one.
That is the quiet irony sitting underneath Ireland's CADA chairmanship. Brussels wants to triple the EU's hyperscale capacity over five to seven years; the member state that already hosts the largest share of it has, out of physical necessity, spent since 2021 constraining new connections rather than expanding them. If CADA's capacity target is to mean anything, most of the tripling has to happen in the Iberian Peninsula, the Nordics, and Central Europe — not in Ireland, and Dublin's Presidency materials do not claim otherwise.
The Sensible Middle Ground
None of this means CADA is wrong to exist. A modest, carefully scoped sovereignty tier for genuinely sensitive public-sector workloads is a defensible hedge against extraterritorial legal risk. But the four-tier assurance scheme should not become a backdoor localization mandate that forces European public bodies to pay a scarcity premium for underdeveloped domestic cloud alternatives, nor should the capacity-tripling target be pursued in ways that ignore the grid and permitting realities Ireland has already lived through. Ireland's own experience — expensive emergency gas generation, a multi-year connection moratorium, and a regulator now requiring new large users to bring their own power — is the most useful case study the EU has for what happens when data-centre growth outruns energy planning. A Presidency that produces nothing more than an honest progress report, grounded in that experience, would be a more useful outcome than a rushed political deal that ignores it.