This article first appeared on GuruFocus.

A potential shakeout could be forming across the global data center sector, as Goodman Group's (GMGSF) chief executive Greg Goodman flagged rising pressure from debt-heavy financing structures tied to the artificial intelligence buildout. Speaking from Sydney, Goodman pointed to billions of dollars in refinancing needs across private equity-backed operators, suggesting that higher borrowing costs could become increasingly difficult to sustain. He indicated that if banks turn more selective in lending, some developers may be pushed toward asset sales, potentially driving consolidation across the industry over the next two to four years.

Warning! GuruFocus has detected 6 Warning Signs with GMGSF. Is GMGSF fairly valued? Test your thesis with our free DCF calculator.

The warning comes as capital continues to pour into AI infrastructure at an unprecedented pace. Data center construction has expanded sharply, rising from about $60 billion in early 2020 to an estimated $340 billion by 2025, while companies are still pursuing large-scale financing to support growth. Bridge Data Centres has engaged banks for a potential loan of up to $6 billion, and DayOne Data Centers is targeting financing as high as $7 billion. At the same time, rising power costs, energy security concerns, and added strain from the Iran war are increasing complexity for lenders, particularly as Asia-Pacific data center investment could reach approximately $800 billion by 2030.

Against this backdrop, Goodman Group is maintaining a more conservative financial approach, with gearing at about 4% compared with an average of 23% among Australian-listed peers. The company is leaning on partnerships to expand capacity without significantly increasing leverage, including a A$3.9 billion collaboration with Canada Pension Plan Investment Board to develop projects across Europe. Goodman expects to have more than A$14 billion of projects underway by the end of June and estimates that around A$150 billion in funding will be required over the next seven to eight years. While the long-term opportunity tied to AI infrastructure could be significant, Goodman suggested that outcomes across the sector may depend on how effectively developers manage capital intensity and evolving demand.

View Comments