'Urban Doom Loop' Author Says Lack Of Computing Power Is Only Thing Holding Back Layoffs

One of the economists who coined the term “urban doom loop” as the office sector reeled from the pandemic has a new warning for landlords: The only thing that's holding back layoffs today is a lack of computing infrastructure.  

The adoption of artificial intelligence will fuel more layoffs as the data centers needed to meet demand from large corporate firms come online, Columbia Business School professor Stijn Van Nieuwerburgh said. 

“In the short run, the demand for the services of these companies is far outstripping the supply of compute,” Van Nieuwerburgh told Bisnow on Wednesday. “Companies can't really do all the work that they want to do with AI yet because there's simply not enough compute to do it, which means that they're not in a position to lay off a substantial fraction of the workforce.”

The high price of computing power today, a function of the number and availability of chips in data centers, is one factor holding back layoffs, but AI has also created a whole new way of working that has yet to be fully realized, Van Nieuwerburgh said. 

The net effect on office demand will be negative, but adoption will be more of a steady flow than a flood as companies opt not to replace employees who leave and make decisions on their office space as long-term leases expire. Widespread impacts will likely take at least five years to work their way through the office sector, Van Nieuwerburgh said.

“In the beginning, there's a real learning curve. This is a new technology that organizations have to adapt to, and that's going to take a long time,” Van Nieuwerburgh said. “It's in that initial phase.”

Van Nieuwerburgh last caught the attention of the real estate world as one of three authors of a 2022 academic paper titled Work From Home and the Office Real Estate Apocalypse, which outlined the concept of an urban doom loop.

The paper outlined how declining office demand in urban cores could lead to falling property values, pulling down property taxes along with it and leaving local governments with less capital to revitalize the area as the quality of life suffers in what can become a cycle. 

Large tech firms are on the leading edge of AI adoption and have been shedding jobs as they pivot internal investment away from headcount and toward massive data center construction.

Meta laid off another 8,000 employees this month, while Amazon started the year with 16,000 layoffs. The tech behemoth is aiming to cut its office footprint by 49,000 desks, which translates to more than 10M SF, this year.

More than 150,000 tech sector employees have lost their jobs this year, according to job loss tracker TrueUp. While smaller tech firms and startups are fueling a rebound in office leasing, the giants are getting leaner. 

Tech leaders are in many ways betting that the cost of AI comes down. A recent internal report that leaked from Microsoft to Fortune showed the tech company was canceling most of its subscription to an AI-based coding tool after heavily promoting it internally and seeing how much its usage cost.

In addition to cost, employee skepticism and unclear productivity improvements are causing executives to rethink AI, Axios reported Thursday.

More data centers are needed to drive down compute costs, but projects across the country face a range of hurdles to development, from securing power to winning community support. 

Data center operators pushed into new markets in recent years, with a Bisnow analysis in March finding that 64% of new data centers were being built outside traditional hubs. 

The push into more populated areas has in many cases been met with steep local opposition, leading at least 25 projects to be abandoned in 2025.