By Timur Tillyaev
Headlines about AI tend to focus on its increasingly astonishing capabilities, concerns about trust, and the race to enact regulation. However, some of the most decisive factors shaping outcomes are far less discussed, namely, the physical requirements to support this technical revolution. Across the continent, the expansion of AI data centres is driving a fundamental reconfiguration of industrial real estate and energy and utility infrastructure.
What is underway is not simply an increase in demand for server space. It is a repricing of land, energy and infrastructure, and a fundamental shift in how long-term capital interacts with Europe’s industrial base. In this sense, AI data centres are less a technology play than a convergence of real assets, energy systems and policy credibility – with real geopolitical benefits for the countries that succeed.
Real-estate investment value
For much of the past decade, data centre investment followed a relatively straightforward pattern. Established hubs offered network access, stable conditions and predictable development pathways. That model is now rapidly changing. The specific power needs of AI data centres mean that grid access, connection delays and permitting friction are significant barriers to growth. Extreme power and significant water requirements for cooling have overtaken connectivity as the primary driver of value.
This requirement for high energy is transforming industrial real estate across Europe. Assets previously seen as straightforward investment opportunities — such as former manufacturing sites — are being reevaluated from a new perspective. Instead of asking “How close is this site to a city or exchange point?” the focus shifts to “How viable is its access to dependable, scalable electricity and water over the next twenty to thirty years?”
As European countries transform their energy grids over the next decade, geopolitics and divergent energy supplies are increasingly key factors in investment decisions. Amid geopolitical volatility, businesses and investors assessing the value of industrial land are becoming more cautious, favouring sites with concrete fundamentals in place. Conversely, assets in markets with rising risks, regardless of any historical fundamentals, are being overlooked. Just look at OpenAI’s recent pause of its multi-billion-pound data centre in the UK, citing concerns about short- and medium-term high energy costs, given the UK’s high exposure to volatility in the energy markets.
A.I.’s specific energy demands
A.I. data centres have very specific energy demands, requiring massive power density (40-120kW per rack) and advanced liquid cooling to support performance, far exceeding the 10-15kW of typical data facilities. This challenges grids designed for more distributed industrial use and pushes energy considerations directly into underwriting decisions.
Large data centres can also use 5 million gallons of water a day, equivalent to the water use of a town of 50,000 people, with a huge impact on the surrounding environment. This means locating new data centres in areas not under water stress or at risk of future water shortages, drought or flooding.
As a result, a new investment model is emerging. Data centre developments are increasingly centred on long-term power purchase agreements, on-site generation, storage and flexibility mechanisms. Energy and utility infrastructure is no longer an external dependency; it is integral to the asset itself.
Returns are being shaped not only by occupancy, but by exposure to energy pricing, grid reliability and regulatory alignment. In effect, investors are underwriting hybrid assets that sit somewhere between real estate, infrastructure and energy — often with durations that extend well beyond traditional development cycles.
The coordination constraint
Europe does not suffer from a lack of capital interested in AI. On the contrary, investors are only too keen to push into new and lucrative transformational technology, with institutional investors, sovereign vehicles and long‑term funds actively deploying eye-watering sums into data centres and energy assets. The binding constraint is coordination.
Planning regimes, grid investment timelines, and permitting processes often have mismatched schedules. When they align, capital flows and projects get off the ground swiftly. When they don’t, projects stall, costs increase, and profits shrink. For investors, the key variable is execution capacity — the ability of regions and regulators to translate their strategies into delivered infrastructure. Markets that can do this consistently are already differentiating themselves.
A structural shift
The growth in AI and the energy intensity that accompanies it is not a short‑term phenomenon. Yes, data centres will make efficiency gains, but AI tools are becoming increasingly complex, and the underlying demand for power-dense infrastructure will only grow.
In many ways, this complexity means AI data centres are closer to transport or energy networks than to traditional commercial property. For investors, that means a laser-disciplined focus on securing the fundamentals: grid resilience, regulatory credibility, community integration and the capacity to scale without political or physical bottlenecks.
Countries that manage energy risk effectively — through diversified generation, resilient grids and credible regulatory frameworks — are positioning themselves as natural locations for AI infrastructure. Europe retains significant advantages in this regard, with deep capital markets, strong industrial capability and growing recognition that digital and energy strategies must converge. Whether these translate into significant long-term investments depends on how effectively each European country treats A.I. infrastructure as a core component of its industrial and energy systems, rather than a standalone technology bet.
The opportunity is substantial. But as with all real-asset transitions, value is not given to those who move fastest, but to those who price and manage the risks correctly.
