Builders of public data centers are fighting the AI ​​boom. But don’t count them out yet

SWith seemingly insatiable demand for computing power, the AI ​​boom has spawned lots of of unicorns from thin air, created dozens of billionaires and created greater than a trillion dollars in market value for major publicly traded technology corporations like Nvidia, Broadcom, Google and Meta. It has also led to an infrastructure land grab that is probably unprecedented in financial terms. Earlier this week, Meta CEO Mark Zuckerberg announced plans to construct tens of gigawatts of AI infrastructure this decade, and “hundreds of gigawatts or more” over time. At $50 billion per gigawatt, that may probably cost trillions.

Boom times for AI infrastructure are more likely to mean glory years for the businesses which have provided it prior to now: data center real estate investment trusts, or REITs. But three of the biggest, Equinix ($78 billion market cap), Digital Realty ($55 billion market cap) and Iron Mountain ($27 billion market cap), don’t see them. Their stock prices have fallen 13%, 11% and 16%, respectively, during the last yr, in comparison with the S&P 500’s 17% rise.

These corporations are the landlords of the Internet; They buy real estate, construct data center shells with supporting infrastructure, and rent the property to technology company tenants. You should kill it. That’s not the case, and that is probably on account of just a few things: lower risk appetite, power constraints, and lack of access to more – and more speculative – capital.

_{Do you could have a tip about AI infrastructure or technology corporations? Contact Phoebe Liu at pliu@forbes.com or phoebe.789.}

Currently, the market assumes that chips from Google, Broadcom, Nvidia and others “will reap the economic benefits of AI, not mercenary data center developers.” [like] REITs,” says Mark Giarelli, analyst at Morningstar. It doesn’t help that the top 20 data center development contracts were won in 2025 from companies other than data center REITs like Equinix and Digital Realty, says Michael Funk, an analyst at Bank of America.

Part of the challenge lies in the REIT structure itself. It was originally developed in the 1960s to provide ordinary people with access to real estate investing by not charging corporate income tax and requiring that each REIT company distribute 90% of its taxable income to shareholders as dividends. Data center REITs emerged at the turn of the century and have largely fulfilled their mission; Since then, their share prices have more or less reflected the market indices. Until AI and its gigantic data centers took over.

Now it’s not working so well. Paying such a high dividend leaves them with less capital to reinvest in an industry where capital comes first. As a result, it’s harder for public REITs to commit upfront to the same colossal amounts of capital that their deep-pocketed competitors tout. That’s a difficult proposition in this new era of data center bidding wars. As a REIT, “you possibly can’t get away like others can,” says Andy Cvengros, co-head of the U.S. data center markets team at listed real estate firm JLL. “It really held her back a lot.”

To land the big deals, Equinix, Digital Realty and Iron Mountain must compete with a seemingly endless list of well-financed companies. Some, like Meta, which both builds its own data centers and rents them from other companies, is funded by cash cow revenue streams. Others, such as private companies Vantage and QTS, both of which build and operate data centers, are increasingly partnering with large investment firms. But REITs’ purchasing power is limited by their risk-averse shareholders, who are more like investors in pension funds than those betting on technology stocks.

REIT shareholders simply don’t want these companies investing in businesses that involve debt with a non-negligible risk of default, which is often what building a data center requires these days. Additionally, REITs often raise less than five times their equity, compared to private companies’ leverage ratios of 10-15.

AAll of this makes it harder for data center REITs to do business with the biggest players like OpenAI – which currently has to pay around $1.4 trillion in expenses with no clear method of payment for it – or Neocloud CoreWeave due to its junk-rated debt.

The difference in risk appetite is perhaps best illustrated by what happened at Oracle in September, when the company announced a profit miss but forecast a massive increase in spending on data center construction. The share price rose 38%. But when Equinix told analysts the same thing in June – cutting revenue expectations but forecasting heavy investments in building AI-focused data centers – its stock price plunged 18% in the following two days.

There are other risks beyond debt. Some multibillion-dollar data center contracts contain provisions such as early termination clauses, which allow data center tenants to terminate a contract earlier than agreed due to wildcard factors such as construction delays. This happens all the time. More than half of data center projects in 2025 experienced delays of three months or more per year JLL. And Microsoft and Meta have each allegedly acted on the premise of those provisions; CoreWeave blamed a delay for its third-quarter profit miss.

Even public data center corporations will not be resistant to the challenges everyone seems to be currently facing when constructing an AI infrastructure. Power bottlenecks are a serious problem because there may be simply not enough grid capability to avoid them. “If we have an empty data center shell that doesn’t deliver megawatts of power, that’s going to weigh on returns,” Giarelli said Forbes. Taken together, this all looks like a nasty deal for a REIT shareholder.

The REITs are cautious in other respects. They have stuck with their core markets like Northern Virginia and Los Angeles rather than buying land in less developed, remote places. But “AI is being built there,” says Funk, because there is space for huge supercomputers. Without more space in already crowded urban areas, REITs cannot grow exponentially. That’s a key reason Equinix’s revenue growth fell short of analyst forecasts last summer, while Digital Realty’s revenue from AI-related workloads fell.

But there is also a chance. Most of AI’s computing power is shifting from building models to executing them, which requires more speed and bandwidth. Data centers in more urban areas are physically closer to the people using AI and also have access to better fiber optic networks than rural data centers. That could give Equinix and Digital Realty a leg up.

Analysts are particularly bullish on Equinix: They say it is best suited for running AI models, in part because of its strong focus on speed and connectivity. The company is already rapidly expanding its xScale business, working with tech giants like Amazon and Oracle as they build data centers like crazy. “We see them changing course pretty quickly,” Cvengros says.

So don’t count out these public data center companies just yet. The demand is so great that there is plenty of room for greater demands in the current boom. Take Applied Digital, once a little-known data center hosting company for crypto miners. In 2023, the focus instead was on building data centers for a few large AI customers. As a result, its stock price and revenue have tripled in the last year, reaching a market cap of $10 billion. But to do that, the company had to take on significant risk and receive a $5 billion capital injection.

The path is more difficult for REITs, which have significantly more to lose, and shareholders who are not keen on a move of this magnitude. They need to ensure that large fluctuations are adequately hedged while raising additional capital. However, to truly benefit from the AI ​​boom, you need to make a bet and commit to it. Otherwise, they will be playing a constant game of catch-up in an AI race where speed is of the essence.

More from Forbes

ForbesCoreWeave’s $29 billion bet that the debt-fueled AI boom won’t failFrom Rashi SrivastavaForbesHow AI is ushering in a new nuclear ageFrom Christopher HelmanForbesThese billionaires have become $450 billion richer by securing AI infrastructure contracts for their own companiesFrom Phoebe Liu