AI Pushing Data Center REITs Toward Big Tech — For Now

The two major publicly traded data center REITs are pouring more money into developing facilities to meet the supercharged artificial intelligence needs of Big Tech, even as they insist their future AI strategy lies elsewhere. 

Equinix and Digital Realty plan to accelerate their pace of development in the coming months, with a growing focus on delivering single-tenant facilities to meet AI-driven demand from hyperscalers like Microsoft, Amazon Web Services and Google.

This kind of hyperscale deployment doesn’t account for the lion’s share of revenue in either firm’s U.S. footprint. They focus on colocation facilities in which servers are leased to a series of smaller users. But both companies indicated hyperscale is the only segment experiencing a measurable boost from artificial intelligence — at least for now. 

The leaders of Equinix and Digital Realty, which have market caps of $80B and $43B, respectively, said on quarterly earnings calls last week that most of their AI growth will ultimately come from their enterprise-focused colocation business lines. These lines are uniquely positioned to connect corporations’ private AI infrastructure with major cloud providers, they said. 

But as corporate AI spending proceeds more slowly than expected, the REITs are driving investment toward the hyperscale opportunity that has already materialized. 

“You're going to continue to see a lot of investment flow to [hyperscale development] … but I think the more differentiated position for us over the long term is unlocking the power of the AI ecosystem through this cloud-adjacent set of offerings,” Equinix CEO Charles Meyers said on the REIT's Wednesday earnings call. “It’s still pretty darn early in the overall cycle.”

Major cloud providers are driving a wave of demand for massive blocks of data center capacity in U.S. markets as they race to secure the infrastructure to support AI, an arms race that has already led to record leasing and a larger-than-ever development pipeline. While a growing number of privately backed developers specialize in delivering exactly this kind of capacity, the two remaining public data center REITs are competing for this turf too. 

Equinix and Digital Realty have U.S. portfolios that are more focused on multitenant colocation, but the two firms have both seen record demand from hyperscalers pursuing AI deployments, and both plan to be even more aggressive in pursuing this kind of development in North America in the months ahead. 

Over the past year, Digital Realty raised over $12B in development capital through asset sales and joint ventures, largely to meet hyperscale demand and build on its existing $6B pipeline. The most significant of these deals was a $7B development JV with investment giant Blackstone, facilitating the build-out of 500 megawatts of data center capacity in Northern Virginia and Europe that explicitly targets hyperscale users. 

“The combination of cloud and AI is driving unprecedented demand for scale and hyperscale capacity,” Digital Realty CEO Andrew Power said on its Thursday earnings call. “Ongoing conversations with customers portend a significant potential acceleration of leasing and development, and we believe we are well positioned.”

Equinix’s appetite for hyperscale development in U.S. markets is more surprising. Until recently, the company steered clear of this kind of project in North America, with its entire JV-funded hyperscale business, branded as xScale, located in Europe, the Middle East and Asia. 

Now, that thinking is changing.

Equinix is planning “accelerated investment” in its overall xScale portfolio, which has 11 projects in various stages of development, with AI and cloud growth providing a “meaningful pipeline of opportunities to drive xScale momentum for quarters to come,” Meyers said. Equinix’s development capital expenditures are expected to rise to as much as $2.8B in 2024, up from $2.6B last year. And that planned expansion now includes the U.S.

Equinix is set to open the initial phase of a Silicon Valley xScale facility this quarter, its first such project in the U.S. And while there are no other xScale projects in the U.S. pipeline, Meyers said Wednesday that it won’t be the last, suggesting that more build-out can be expected in the months ahead. 

“In terms of U.S. xScale, we are absolutely working on how we're going to continue to be more aggressive in this market. We think there is opportunity,” Meyers said. “Our tune, and my tune specifically, has changed a bit on that over the last couple of years.” 

But while hyperscale may be the only segment where data center REITs are feeling AI impact their bottom lines right now, Equinix and Digital Realty anticipate that their enterprise-focused colocation businesses will ultimately be the biggest beneficiaries of the AI revolution. The two firms have similar beliefs that their distinct data center portfolios leave them uniquely positioned to capture a key chunk of the emerging AI marketplace. 

The major cloud providers like Microsoft and Google have driven corporate AI adoption, and cloud-based AI models will likely continue to account for a significant percentage of corporate spending on AI. But many companies will want to control some or all of their AI computing themselves.

These could be smaller tech firms developing their own AI models. More likely, they will be corporations using cloud-based AI, but because of security, compliance or economic reasons, they will need to have more control over some of their data than cloud providers can promise. 

Called hybrid AI infrastructure models, these are deployed by companies like pharmaceutical firms applying generative AI to sensitive, proprietary data that they want to store themselves, financial firms navigating strict compliance rules, and companies handling government data that needs to be processed or stored in specific jurisdictions. 

“The focus is on where people want to place their data. Sometimes it’s about the proprietary nature of that data and controlling it, and sometimes it’s about the cost of moving data in and out of public clouds or factors like performance,” Meyers said. “People are saying, ‘I want to maintain my control over my enterprise data, and I want to place it somewhere that is cloud-adjacent because the hyperscalers are innovating at such a rapid rate that I want to use their models and their tools.’” 

Typically, these firms will want their own computing infrastructure hosted in facilities close to the major cloud providers where much of the AI computing is taking place. With portfolios skewed toward massive amounts of capacity in major metros with strong connectivity to major cloud facilities, the leaders of both REITs said they are confident that they are uniquely positioned to control a significant share of the market for private and hybrid AI. 

“[We’re] hitting in that sweet spot of what we think customers are really looking for: control over their enterprise data, the ability to access AI tools from the hyperscalers who are innovating rapidly in that area and stitch it all together and make it work in a way that makes sense for them,” Meyers said. 

He said he expects AI spending to mirror the overall cloud market, in which more than half of enterprise customers operate their own information technology infrastructure or deploy a hybrid model. It is a market that many across the industry believe is emerging quickly. While just 5% of enterprise data center customers used generative AI at the beginning of last year, that number is expected to leap to 80% by 2026, according to a Gartner study. 

Equinix and Digital Realty are making significant investments based on this belief, launching colocation products where they provide not just space and power but also the high-performance computing equipment tenants need for their AI workloads. These processors are expensive and hard to acquire, and providing them gives tenants the ability to deploy quickly in an AI landscape where speed is everything amid a wave of innovation. 

“We launched our offering last year around the high-density colo in anticipation of a lot of private AI type of deployment coming to market,” Digital Realty Chief Technology Officer Christopher Sharp said on the earnings call. “We definitely see the long tail of that value happening in inference and private AI. … It's something you'll see play out over this year.”

This AI-driven demand for colocation product is emerging more slowly than the leadership of the two data center REITs had hoped.

For Digital Realty, this was a contributing factor to quarterly earnings figures that were at the lower end of investors’ expectations. The firm reported quarterly revenue of $1.4B, up 11% year-over-year but short of projections, resulting in a more than 7% drop in the company’s share price.

Equinix’s quarterly results were stronger, with $2.1B in quarterly revenue marking a 15% year-over-year increase. Still, the company’s leaders said demand for its retail colocation product from tenants looking to deploy AI workloads was lower than they had anticipated.

“I wouldn't say it's yet proving to inflect our retail bookings. We were a little shorter than we wanted to be, but we're seeing the green shoots there,” Meyers said. “We continue to be very optimistic about that, but, tempering expectations, it's going to take time for that to really fully realize itself in terms of bookings.”