Executive Summary
- The sector is attracting interest and capital from a broad spectrum of investor types – with some seeing data centres as an opportunity they can’t afford to miss.
- But, new research has identified critical expertise gaps that may leave investors vulnerable to risks that they don’t fully understand.
- Iryna Pylypchuk, Director of Research at INREV argues that for some, it is more likely to be a risk they can’t afford to take.
As AI, cloud computing, and the wider digital economy drive demand for digital infrastructure, institutional real estate investors are examining their exposure to data centres, drawn by the sector’s growth potential.
New research from INREV – the leading industry body for institutional investors in private real estate – identifies critical expertise gaps that may leave investors vulnerable to risks that they don’t fully understand.
The paper, titled ‘Data centres: navigating a fluid asset class’, draws on in-depth expert interviews with investors, fund managers, developers, and operators. They collectively represent over €1 trillion (£860 billion) in total alternatives assets under management (AUM). It finds that data centres do not fit neatly into any single asset classification but are an inherently hybrid asset type that feature a mix of real estate, infrastructure and private equity characteristics.
The sector is attracting interest and capital from a broad spectrum of investor types. But few investors house the full breadth of expertise needed to understand all the associated risks under one roof because these are quite varied. They could include, but are not limited to, power and grid connectivity, environmental impacts, planning constraints, fast technological change, social and governmental resistance, and tenant quality.
Five principal investment models
The study identifies five principal investment models that investors are adopting when entering the data centre sector. These range from early-stage entry points, such as buying undeveloped land which has secured planning consent and grid connection, through to long-term, direct ownership and operation of a stabilised, income-generating data centre asset.
- ‘Powered land’ – where investors buy land which has secured planning consent and grid connection
- ‘Build-fill-sell’ – where investors build the data centre facility and secure tenants before selling
- ‘Asset repositioning’ – repurposing existing assets into a data centre or a data-centre ready site
- ‘Core open-end fund strategies’ – long-term ownership of a stabilised, income-generating data centre asset
- ‘Operational platform ownership’ – combining development capability with direct operational expertise
Different models are being adopted to reduce risk and enable investors to invest in data centres at the appropriate point in the value chain for their specific specialisms and knowledge. For example, the technical expertise required to invest in a data centre as a long-term owner-operator may not be required to buy undeveloped land or pursue a buy-develop-sell strategy.
Risks are complex and wide-ranging
The paper ultimately issues caution to investors in the real estate sphere looking to capitalise on the data centre boom. The asset class carries risks with no real equivalent in traditional real estate, and assessing risk profiles requires fluency in highly technical areas spanning power markets, grid connections, technical specifications, and planning.
Overall, the paper finds that deploying institutional levels of capital towards data centres may not be suitable for all investors. It requires a combination of substantial capital, specialist expertise across a wide range of areas, and the ability to navigate a complex cross-section of power and infrastructure, technology risks, and political and social considerations.
Iryna Pylypchuk, Director of Research at INREV, said: “There is no doubt a role for institutional capital to play in powering growth for the data centre sector, which is quickly cementing itself as the ‘critical infrastructure’ of the future. And that critical infrastructure is not anchored on AI alone. It’s also about the wider adoption of cloud services and digitalisation in our lives, at home and at work. However, investors must not rush for the sake of rushing.
“They should resist treating data centres as an asset class that every institutional portfolio should necessarily hold. There is a danger that investors see data centres as an opportunity they can’t afford to miss, when for some, it is more likely to be a risk they can’t afford to take. The case for any allocation remains investor specific, with the decision ultimately depending on whether data centres fit the portfolio and whether the investor is equipped to underwrite and manage the sector’s distinct risks.
“As the European data centre market evolves and institutional participation continues to broaden, investment decisions will become less about securing exposure to a high-growth sector with strong structural tailwinds, and more about determining the appropriate scale and route of that exposure.”
