Data centres

Answer

Data centres have emerged as a distinct commercial real estate asset class, driven by developments in cloud computing, artificial intelligence (AI) and general growth in digital transformation. There’s high demand meaning that the sector is booming from an occupational perspective, making it increasingly attractive for both developers and investors seeking exposure to the digital infrastructure that underpins the modern economy. The sector isn’t without its challenges as power demands, location, environmental considerations, and technology and engineering requirements, among other factors, all have a part to play in data centre projects. These intersecting constraints, combined with unprecedented demand and limited supply, have created intense competition among developers, investors and operators. This adds to the buzz around data centres. As a recently emerged asset class, data centres are facing issues relating to investment and their financing. This article will explore some of these challenges and nuances of the data centre sector in further detail.

Power

Data centres are highly power intensive. Power has now become the dominant requirement in data centre development, eclipsing fibre connectivity, which previously shared equal importance. Within the UK, demand for data centre grid capacity in high-density Tier 1 urban locations has a connection queue timescale of six to eight years. The UK has reformed the grid connection process, hoping to prioritise the most viable projects. National Grid is implementing measures to remove speculative projects from the queue, with anticipated regulatory changes to require demonstrable land rights and readiness to take power. This adds to a 'use-it-or-lose-it' mechanism to ensure the existing grid capacity is being used immediately by developers.

While the preference from customers remains safe, regulated grid connections, the queue pressures have prompted exploration of alternative power solutions. In the US market, gas turbines are commonly used as backup, although this approach faces resistance in the UK from clients with sustainability concerns. Some operators are considering biomethane, while certain markets (eg, Poland) are exploring small nuclear reactors; however these remain several years from viability. Each alternative presents significant regulatory complexity, design challenges and capital risk. Hyperscalers have the option of building their own microgrids, creating planning issues and contractual negotiations with local communities and easement owners over the land, although the regulatory uncertainties and high capital requirements mean such solutions typically require very substantial demand in a specific region.

Land and location

Certain jurisdictions have plenty of space and are seeing massive investment (notably the US and Australia) in the data centre industry. Conversely, locations within Europe that need data centre capacity and have the power supply are lacking in space; Amsterdam has already banned the building of further data centres until 2035 to avoid overcrowding. This is pushing companies to compete against each other to identify and develop the best locations for data centres. While the Scandinavian countries have an abundance of space, they lack the connectivity and power supply that’s necessary for data centres. However, new markets are emerging to fill this gap, with Zaragoza in Spain (benefiting from green energy availability) and Sines in Portugal becoming increasingly attractive locations for data centre development.

Latency has also become an increasingly important factor in site selection. For applications requiring real-time processing, such as financial trading platforms and autonomous vehicles, even milliseconds of delay can be commercially catastrophic. This has driven demand for edge data centres located closer to population centres and end users, intensifying competition in the limited number of locations that have the required latency and power supply. Consequently, data centre developers must carefully weigh the trade-offs between locating facilities near urban centres (minimising latency but facing higher land costs, power constraints and planning resistance) versus more remote locations (offering cheaper land and better power availability but potentially compromising performance for latency sensitive applications). This consideration increasingly influences contractual negotiations, with hyperscalers and operators demanding specific latency guarantees in their service level agreements, creating new elements of risk allocation issues within lease documentation.

Planning, social and environmental factors

Beyond power and land constraints, data centre developers face increasing regulatory and community resistance. There’s growing concern about the extreme rate at which data centres are being built, with questions raised about their actual job creation potential and impact on local energy grids. While developers often promote employment benefits, critics argue that operational data centres require minimal staffing compared to traditional industrial developments of similar scale. Environmental groups have also increasingly challenged data centre planning applications on grounds of water consumption (particularly for cooling) and carbon emissions. Some jurisdictions (notably the EU) are responding by imposing stricter environmental and regulatory conditions, requiring developers to demonstrate carbon neutrality plans or commit to renewable energy sourcing before granting planning permission.

Technology and engineering considerations

The construction requirements of data centres are changing as technology grows. For example, cooling systems are changing from air-powered to water-powered systems and this is rapidly influencing the commercial considerations underpinning site selection, with proximity to water sources and sustainable drainage systems becoming critical factors in feasibility assessments. The shift to liquid cooling technologies presents both opportunities and challenges for developers. While more energy efficient than traditional air cooling, liquid cooling systems require different infrastructure specifications, potentially increasing upfront capital costs, while reducing long-term operational expenses. This creates tension in negotiations between developers and operators regarding who bears the cost of these technological upgrades and how such investments are reflected in lease pricing or profit-sharing arrangements. This can be particularly pertinent in light of tight construction timeframes often imposed by end users.

The competitive pressure is intensified by the massively tight timelines being imposed by end users. The market is currently undersaturated with low vacancy rates driving a race among developers to deliver operational capacity ahead of competitors. Companies need to keep ahead of the pack to ensure that the significant upfront cost of a data centre development remains commercially viable in a market where being second to deliver in a particular location can mean the difference between full occupancy and struggling to attract tenants. This time pressure creates legal challenges around acceleration provisions, liquidated damages for delay, and the allocation of risk for unforeseen technical or regulatory obstacles that could derail tight construction programmes.

Market dynamics

The data centre sector presents notable barriers to entry and exit, with significant capital investments representing sunk costs that can’t be fully recovered upon exit, making this an illiquid asset class. The current market is characterised by a supply-demand imbalance, with real estate vacancies falling across major markets: the Big 5 markets all report sub-10% vacancy rates (some below 5%), while the US overall has a sub-2% vacancy rate. This scarcity has driven premium valuations for data centres, with data centre real estate investment trusts (REITs) currently trading at a premium. Prices are rising, necessitating greater risk consciousness from investors entering the market. However, there’s a legitimate question as to whether data centres are becoming overhyped despite their current strong fundamentals.

Financing challenges

Securing, building and running data centres is likely to require substantial investment. The financing environment presents several challenges that intensify competition for viable projects. The degree of AI-related spending by borrowers and the sustainability of this outlay, together with the risks of investing in quickly depreciating assets, are likely to raise concerns for lenders when considering investing in this area. The current interest rate environment means investors may demand risk premiums for their equity, while the lack of core capital further constrains the market. Exit opportunities remain limited, with relatively few data centre exits occurring to date, raising liquidity concerns for investors.

Valuing data centre assets presents additional complexity for both lenders and investors. Unlike traditional commercial real estate, data centres can’t be assessed purely on location and square footage. Their value often hinges on the income potential of the data centre, which will include connectivity capabilities, available infrastructure (eg, power and cooling) and technical specifications. This means that financing arrangements must account for greater uncertainty around asset valuation, particularly for newer developments without established operational track records. Lenders increasingly require detailed technical assessments from a valuer with expertise in this area alongside standard property valuations, adding time and cost to the financing process.

The emergence of debt funds and alternative lenders has partially addressed the financing gap left by traditional banks' cautious approach. These alternative capital providers often accept higher leverage and construction risk in exchange for enhanced returns, although their involvement may introduce additional legal complexity around intercreditor arrangements and enforcement rights.

Market evolution

The private equity route for funding may provide opportunities in this area. Some co-location operators are considering initial public offerings and, in particular, conversion to a REIT. While several listed REITs exist, there’s a notable absence of broader fund vehicles. This presents an opportunity for fund managers to develop new investment products tailored to the data centre sector, which will in turn create work for legal teams in terms of fund formation and structuring. Institutional capital groups with low costs of capital haven’t yet entered the space in force; when they do, their entry will likely coincide with a lower interest rate environment, potentially reshaping the competitive landscape. US hyperscalers maintain massive buying power, further influencing market dynamics and pricing. Their procurement strategies and contracting requirements often set the standard for the broader market.

Final thoughts

The buzz around data centres is more than hype but rather reflects a fundamental shift in how digital infrastructure is valued and financed. The sector's challenges are structural features that will continue to define the market. Power will remain the critical constraint, suitable locations will grow scarcer, environmental scrutiny will intensify and technological requirements will continue to evolve. Institutional capital will enter the market as interest rates normalise, while the requirements of a data centre itself will change in line with technological advances and regulatory environments. These dynamics ensure that data centres as an asset will remain one of the most challenging and commercially significant areas of real estate practice, requiring advisers who understand both the technical complexities and the commercial imperatives shaping client decision making in this rapidly evolving sector.

Georgina Low is a trainee at Taylor Wessing.