Dublin’s data centre dilemma as Europe’s AI gold rush meets grid reality
NEW FOR MIPIM 2026: Data Centers Summit – From Bits to Bricks
At MIPIM 2026, the language around data centres was bullish, but the message beneath it was more sober. Europe is in the middle of a major expansion cycle driven by AI, cloud computing, 5G and sovereign data ambitions. Capital is flowing, new markets are emerging and investors are scanning the continent for opportunity. Yet, behind the optimism, is the knowledge that the sector’s future sucess will be determined by planning and access to power.
Ireland, and particularly Dublin, was singled out as a region previously desired, but now congested and not in line for growth.
The Data Centers Summit made clear that the core European markets remain deeply attractive, but they are also “massively constrained.” Access to grid connections was repeatedly identified as the single greatest obstacle to development. In market after market, the wait for power is long, uncertain or simply impossible within commercially acceptable timelines. That is pushing investors toward regions such as Iberia and the Nordics, where power is cheaper, more available and, crucially, potentially deliverable before 2030.
Dublin sits at the centre of that tension. It remains one of Europe’s established digital infrastructure hubs, but it was cited during the panel as an example of a market where power constraints have become entrenched. One panellist referred to a moratorium in Dublin that has been in place for several years and described how operators are already relying on alternative solutions to keep projects moving. It was uncomfortable to hear. In a city that has become strategically important to global digital infrastructure, the grid is no longer simply a utility issue. It is now a development constraint, a planning constraint and, increasingly, a policy test.
The panel discussion hinted at a broader shift in the market. This is no longer just about securing land and then expecting infrastructure to follow. A viable data centre site now needs what one speaker described as every ingredient in place: power, permitting (planning), environmental clarity, zoning, labour supply and a realistic delivery timeline. “Oven-ready” and de-risked sites are attracting the strongest buyer interest. Everything else is moving into a more speculative category.
That is where the summit’s most revealing phrase emerged: this is the “new gold rush.”
In one sense, the comparison is understandable. The sector is drawing institutional capital, sovereign wealth, infrastructure funds and logistics developers. Industrial landowners, infrastructure investors and traditional real estate players are all looking at data centres as the next high-value asset class. The attraction is obvious: AI demand is rising, digital infrastructure is increasingly treated as critical national infrastructure, and a site with credible power access can command a very significant premium.
But the panellists also made clear that gold rushes create distortions.
The most pointed warning came in the discussion of “zombie, old ghost projects.” These are schemes that absorb attention, grid applications, supply chain bandwidth and speculative capital, yet are unlikely ever to be built. The panel cited the UK as an example, where more than 200 applications linked to AI growth zone ambitions may never proceed because they lack business case, customer demand or policy backing. The critique was not that ambition is unwelcome. It was that a surge of poorly grounded projects can clog the system, inflate land expectations, distort pricing and create false signals for the market.
Wherever infrastructure is constrained and investor appetite is strong, the temptation grows to treat any site with land and a notional power connection as a future data centre campus. But the summit repeatedly challenged that assumption. A powered land parcel is not, by itself, a bankable project. It takes deep capital, operator capability, planning readiness, resilience design, customer alignment and long-lead supply chain commitment. One panellist noted that developers may be tens – or even hundresd – of millions deep before a site is even ready for market. In other words, the distance between a land story and a deliverable scheme is much greater than many promoters would like to admit.
For Dublin and Ireland, this is litmus test The real issue is not whether the country remains attractive to data centre investors. It does. The issue is whether policy, grid management and planning processes can distinguish between strategic projects and speculative noise.
That has implications well beyond the data centre sector. If grid access is scarce, governments and regulators are forced into difficult prioritisation decisions. Panellists referenced the tension explicitly: if capacity is limited, how should policymakers rank data centres against hospitals, factories, housing-linked infrastructure or defence-related investments? The old first-come, first-served model is beginning to look inadequate in an era of structural constraint. Several speakers pointed to reforms elsewhere in Europe that attempt to filter speculative applications and attach more discipline to grid requests.
Ireland may need to confront similar questions more directly.
Dublin’s challenge is not simply that demand is high. It is that the city has become a collision point between digital infrastructure ambition, finite grid capacity, public concern and competing national priorities. That means the policy response cannot be binary. It cannot be a blanket hostility to data centres, nor can it be an open invitation to every speculative proposal dressed in AI language. The market is already signalling a more mature standard. Certainty, not noise, is what capital values.
There is also a deeper community and stakholder engagement question. Several speakers stressed that data centres are still poorly understood by the public and often encounter reflexive opposition. Misconceptions persist around water use, local benefit and community impact. Yet if AI, cloud services and digital sovereignty are now essential parts of economic and public life, then the physical infrastructure behind them has to be planned as part of national development rather than treated as an afterthought. That requires more honest public engagement, more transparent infrastructure planning and stronger alignment between national policy and local implementation.
For Ireland, that means acknowledging two things at once.
First, Dublin’s position in Europe’s digital economy is real and valuable. It should not be casually weakened through drift, opacity or indecision. Second, the era of speculative exuberance is already producing collateral damage. Ghost projects and grid queue gaming do not represent growth. They represent system failure.
The summit’s most important takeaway was a simple one: certainty wins. In a power-constrained market, the projects that survive will be those that are properly underwritten, credibly permitted, realistically powered and clearly connected to demand. The rest may generate headlines, valuations and conference chatter, but not infrastructure.
