A version of this article first appeared in the CNBC Property Play newsletter with Diana Olick. Property Play covers new and evolving opportunities for the real estate investor, from individuals to venture capitalists, private equity funds, family offices, institutional investors and large public companies. Sign up to receive future editions, straight to your inbox.
As hyperscalers like Nvidia, Amazon, Google and Meta announce more and more data center projects, cries of a bubble have been growing. Some say the sector is already overbuilding for a market that is still in its infancy with many unknowns ahead. There are also concerns that the financing for some of these projects is risky.
Andy Power, CEO of Digital Realty, the second-largest data center REIT in the world, says just the opposite.
Power has been working at the company for 25 years and said he is not concerned about too much construction in the sector.
“Based on the actual real demand from real customers with real long-term,15-year contracts, we are not in an oversupply state today,” he told Property Play.
The global data center sector is poised for continued unprecedented expansion, with capacity expected to nearly double from 103 gigawatts to 200 gigawatts by 2030, according to a new outlook from JLL. That is being driven, of course, by artificial intelligence, which JLL says is rapidly reshaping the data center landscape. The real estate research firm forecasts that AI workloads will represent half of all data center capacity by 2030. It also says that “property metrics do not point to a bubble.”
In fact, JLL predicts that sector growth will need up to $3 trillion in total investment over the next five years, including $1.2 trillion in real estate asset value creation and approximately $870 billion in new debt financing. The report calls it an infrastructure supercycle.
“We’re witnessing the most significant transformation in data center infrastructure since the original cloud migration,” said Matt Landek, global division president for data centers and critical environments, at JLL. “The sheer scale of demand is extraordinary. Hyperscalers are allocating $1 trillion for data center spend between 2024 and 2026 alone, while supply constraints and four-year grid connection delays are creating a perfect storm that’s fundamentally reshaping how we approach development, energy sourcing and market strategy.”
JLL forecasts AI workloads could represent half of all data center capacity by 2030, compared to approximately 25% in 2025.
The outlook from Digital Realty’s Power is more fundamental. He says the sector is simply building on technology trends like cloud computing and digital transformation that have a long tailwind.
“Will there be ups and downs along the way? I’m sure there will be,” Power said. “But these are trillion-dollar companies that have real cash flow and businesses that are investing for this innovation. And we in digital and data centers, the way we do it especially, are really trying to do that in a way that is long-term durability that will insulate us and help cater to all those in that region.”
Power also said that the real estate side of the AI arms race is at less risk than the hyperscalers themselves.
“In our strategy and the bricks and sticks and physical infrastructure we invest in, I see tremendous insulation towards any type of shock. We are essentially in a place where demand is well outpacing supply, so the speculative data center builds, you can’t build it fast enough for the customers,” Power said, adding that vacancies at Digital Realty are the tightest they’ve ever been.
As with all real estate, Power also pointed to location. Digital Realty is investing in areas where workloads necessitate data, like Northern Virginia; Chicago; Dallas, Texas; and even Singapore, Tokyo, Frankfurt and London – “proximate to the eyeballs, the consumption, the devices,” he said.
On the financing side, however, Starwood Capital Group chairman Barry Sternlicht and others have raised concerns.
“What we’re watching now is the credit worthiness of the tenant, and particularly Oracle, because Oracle is doing all these deals backended to Chat[GPT],” Sternlicht said on the Property Play podcast in November. “And Chat is a startup that doesn’t make money and requires hundreds of billions of dollars to grow to the scale that they want to be.”
Power noted that all the companies involved, Oracle included, have tremendous businesses outside of AI and (with the exception of Oracle), they all want to own their real estate. As of now, for data centers, they own about half.
“They don’t believe that they’re going to be walking away from these leads in the markets,” he said.
