News
Data centers reshape real estate as banks struggle with funding
The Stargate AI data center under construction in Abilene, Texas, US, on Tuesday, Sept. 23, 2025. Stargate is a collaboration of OpenAI, Oracle and SoftBank, with promotional support from President Donald Trump, to build data centers and other infrastructure for artificial intelligence throughout the US.
Kyle Grillot | Bloomberg | Getty Images
For decades, global real estate revolved around what could be seen — gleaming office towers, shopping malls, and trophy hotels, but that has increasingly shifted to “invisible” property: cloud and data centers.
“The world of real estate is changing from what I’d call the ‘visible’ to the ‘invisible,’” said Kishore Moorjani, CEO of CapitaLand Investment’s alternatives and private funds team.
“[It’s] everything we can’t see but we use: the cloud, as we like calling it, lives in the data center,” he noted during a panel at the Milken Institute Asia Summit held in Singapore. “That’s what’s driving the value shift.”
According to a recent 2025 data center investor interest survey by real estate services firm CBRE, 95% of major investors worldwide plan to increase their investments in data centers. Of the 92 major investors polled, 41% said that they planned to allocate $500 million or more in equity to the data center sector in 2025, compared to 30% last year.
Data center demand has surged in recent years, largely driven by the explosion in AI workloads, which require vast computing power, electrical power, cooling and networking infrastructure. Goldman Sachs forecasts that global power demand from data centers will rise 50% by 2027 and by as much as 165% by 2030.
Stuart Crow, CEO of APAC capital markets at global real estate firm JLL, said investors are reallocating portfolios away from traditional sectors and into alternatives.
“A big part of that is data centers, and now battery storage and infrastructure that’s associated with that,” he said at the same panel.
Goodwin Gaw, managing principal at Gaw Capital Partners, another panelist, added that his firm is investing in China’s data centers.
Capital constraints?
However, a funding gap is emerging, the panelists noted. While global data center appetite is strong, the pace and scale of these projects are straining traditional lenders, said the panelists.
“Banks are certainly challenged around their data center exposures, given the sheer volume and quantum of the build that’s going on, what it costs to put that capacity in place,” said CapitaLand Investment’s Moorjani.
A hyperscale facility — which is a massive data center — costs $12 million per megawatt to build, and modern hyperscale data centers range from 150 to 300 megawatts, according to Columbia Business School. AI-centric facilities that exceed 1 gigawatt in capacity will require multi-billion-dollar investments.
A gigawatt is 1,000 megawatts of power, while a megawatt is 1 million watts.
“I think increasingly the banks are starting to feel a little bit strained,” Moorjani added.
Boston Consulting Group estimates that hyperscalers will need to invest roughly $1.8 trillion between 2024 and 2030 to meet AI and cloud demand.
“So the biggest question mark for the real estate community is: is there enough capital at the moment?” said JLL’s Crow.
The panelists also noted that the rise of artificial intelligence could reshape physical real estate, particularly office space, as companies reduce headcount and adjust their office footprint needs.
Property consultancy Savills, in a report published September, forecasted a slower rebound in global real estate investment this year. Global commercial real estate investment is expected to rise 8% in 2025 from last year, well below the 27% previously forecasted by the firm.
According to the report, many developers face obstacles such as rising construction costs, financing constraints, labor shortages, and regulatory complexity. Still, large pools of undeployed capital remain available, and major institutional players show little sign of abandoning real estate as a core asset class.
