In the evolving hierarchy of commercial real estate, data centers have become a top-tier property type for many institutional investors, joining logistics and housing atop their wish lists.
Within the broader category of alternative real estate, digital assets were far and away the most favored, in the PERE Investor Perspectives Study 2022, with more than 60 percent of respondents targeting the property type. The next closest niche was healthcare real estate at 15 percent.
Driving this demand is the industry’s stellar performance during the pandemic. Listed data centers in the US delivered total returns of 21 and 25 percent, respectively, in the last two years, according to trade group Nareit. With top markets enjoying occupancy of more than 90 percent amid record deliveries of new product, there is no reason to doubt the sustainability of that success.
Beyond the numbers is the idea that data centers represent a future-proof way to build exposure to real estate. In the wake of major disruption to offices and retail centers, once the bedrocks of institutional property portfolios, this may be a comforting thought for investors. But it does not reflect the whole reality on the ground: just like work and shopping places, data centers are facing obsolescence challenges too.
This week, managers Harrison Street Real Estate Capital and American Real Estate Partners committed $1 billion to develop data center campuses in Northern Virginia, the world’s biggest hub for such facilities. The impetus behind the strategy is a dearth of suitable data centers for hyperscale cloud computing providers. Despite being home to 31 million square feet with another 51.5 million square feet under construction, according to an H1 2021 market report from brokerage JLL, the region lacks properties with enough power capacity for groups such as Amazon Web Services, Microsoft Azure and the Google Cloud Platform, the managers say.
To address this, Harrison Street and AREP are constructing multiple data centers each on two parcels of land. Doing this makes it easier for local utilities to increase energy availability through onsite electrical substations. That strategy is not easily implemented for other data centers, which tend to have been built one to a property and with much smaller power needs in mind.
If data centers built less than 10 years ago are already ill-equipped to meet the needs of today’s top tenants, that does not bode well for the future relevance of the current stock.
For would-be owners, this means they are better off building their way into the asset class rather than buying, and their success hinges on both securing a large enough plot of land for multiple centers and bringing in supplemental electricity. Those are high barriers to entry.
The Harrison Street-AREP joint venture highlights that data center growth is ultimately a technology trend, which puts it in jeopardy of being disrupted by the next big breakthrough. That could mean a surge in power demands that surpasses the capabilities of even the newest centers, or an innovation that allows data to be stored so efficiently that data centers are no longer necessary. Already there is talk of a widescale swapping out of terrestrial locations for satellites, for example.
For now, the performance and growth prospects of data centers make them worthy of institutional ownership. But if investors feel this is a property type immune to the sorts of disruption visible elsewhere, they should reconsider that view.