February 5, 2025 - Like any form of investing, real estate investing relies on getting ahead of market trends. One rising trend that is difficult to ignore is the emergence and rapid growth of data centers. Data centers are large-scale facilities used to store critical applications and data. They are used for cloud storage, mobile network traffic, and overall internet traffic. Soon, data centers will even be a key component for the rise in AI, which will be the next big change in American industry as we strive to make operations leaner and more efficient.

According to Colliers 2024 U.S. Capital Markets report, with demand for data centers projected to grow around 10%-20% annually until 2030, an increasing number of investors, developers, and new market players are taking notice. However, important aspects should be considered before investing in the data center space.

Per PWC’s 2025 Emerging Trends in Real Estate, data centers have been the most sought-after investment for two years due to the high demand for the services this asset class provides. This is not surprising, as mentioned above, data centers are used for cloud storage and for internet traffic.

While demand for data centers has skyrocketed, supply has been constrained, primarily due to issues associated with limited power transmission capabilities. Due to power and other unique requirements of data centers, data centers are concentrated in certain areas of the United States. In fact, approximately 70% of the world’s internet traffic passes through 300 data centers in Northern Virginia. Key characteristics of a desirable data center site include low land costs, power availability and cost, fiber connectivity, government support, proximity to densely populated areas, and low risk of disaster or disruption.

Types of data centers

There are three primary types of data centers:

• Enterprise facilities: facilities that are built, maintained, operated, and managed by companies for the optimal operation of their technological equipment.
• Cloud or hyperscale: facilities that are built to accommodate the scalable applications revolving around the cloud, big data, or distributed storage. It is noted that there are many hyperscale data centers that cater to individual companies with large mobile traffic, internet traffic, and cloud storage requirements. The benefit of owning hyperscale facilities is that the leases last 25+ years, and it is highly likely tenants will renew, given that the location and the IT infrastructure of the data center are key to them running their business.
• Multitenant, colocation, or network dense: these dense facilities are built to accommodate multiple companies that lease space within the data center. Network dense facilities allow users to interconnect with many different network service providers, enhancing the interconnectivity of the user to other facilities within a market and outside that market.

The explosive growth in the data center market is expected to continue as more data is required for cloud, mobile, and internet usage over the next five years. It is expected that global data center capacity will double within that time for uses such as the increase in digitization of banking and financial services transactions to the expansion of usage of social networks. With the guarantee of increased use within the next five years in mind, 97% of investors plan to allocate more funding to data centers in 2024, according to CBRE.

It is expected that the most significant growth in data centers will be due to AI. As many corporations look to AI to more efficiently sort through data, allocate resources based on algorithms, automate repetitive tasks, and use data based on consumer patterns for better marketing campaigns, they will need data centers to provide the resources, storage, and networking that AI applications require to function.

It will be important that these facilities will be much larger than the facilities we have now as AI usage will require high computational power and very low latency. With this in mind, to make sure that they have the storage capacity necessary for future AI-related projects, many large corporations, such as Microsoft, Blackrock, and even the US government, are willing to pay billions to get data center facilities operational so that they can satisfy these storage capacity requirements.

Barriers to data center entry

There are three notable barriers to data center entry. The first is that the most profitable and popular data center type, hyperscale facilities, can be expensive to construct and costly to purchase the necessary equipment to operate. However, many developers have partnered with REITs (real estate investment trusts) to get the capital to develop hyperscale centers. Even with completed hyperscale facilities, there is often an option to co-invest alongside deep-pocketed lead investors such as REITs, as many of the owners of these data centers will need additional capital if they want to expand their facilities or update their equipment.

The second barrier is that these data center facilities require substantial power to operate. The current electrical grid will not be able to satisfy the substantial power requirements that data centers will need once AI is more widely used. AI uses three to four times the amount of energy than storing data for the internet does, and if another option is not explored, there is a likelihood of brownouts. However, many in the data center sphere have found a resolution to this barrier – investments in nuclear power. Nuclear power can produce 8,000 times more energy than traditional fossil fuels due to the high energy density of nuclear fission, making it considerably more efficient in terms of energy production. Nuclear energy is not only very reliable and can generate more energy than traditional fossil fuels but it is also carbon-free.

The third barrier to entry is that these facilities require massive amounts of capital spent on cooling systems and can deplete an already thin water supply. Like the last barrier, companies have implemented creative ways to address this issue and will continue to develop new approaches.

Investment approaches

Investors seeking to invest in data centers generally can choose from the following potential approaches:

• Directly investing in stabilized or near stabilized data centers: This approach is most concentrated among large operators and big technology firms.
• Developing data centers: This approach can take different forms with the investor either acquiring land, completing entitlements, securing power, and handing over a shovel-ready site, or build to suit development, or speculative development.
• Acquiring a minority or controlling stake in a data center platform: This approach allows property developers to purchase interest in already established data centers. Many that own data centers would welcome this additional capital to expand their facilities and purchase equipment to make sure the facility is operational.

While some significant barriers to entry exist and supply is well short of demand, the consensus is that interest in data centers will continue to grow and provide outsized returns for the foreseeable future.

This is virtually a sure thing given the expected rise in internet data usage and the continued emergence of AI, which by 2030 is projected to attract annual investments in excess of $800 billion. Given the results to date and the extremely positive outlook for the future, real estate investors should be aware of this asset class and consider opportunities as they arise.

Get in touch with our real estate practice.