The data center boom has made it clear that computers will need places to work. What about the humans?
Large tech firms have been laying off workers as they plow their capital into massive data facilities. In mid-May, Meta downsized by 8,000 employees just after raising its forecast for 2026 spending on data centers by 8 percent, to a range of $125 to $145 billion, according to its first-quarter earnings.
Although the tech behemoths have so far not given up major office space, none have been obviously expanding either. Earlier this year, Meta did sign a retail lease for a New York City shop for its wearables, and Google leased a building in Washington, D.C., near Union Station. But Google also made plans to lease space at its upcoming Thompson Center redevelopment in Chicago to other firms.
For commercial real estate, the future depends on more companies than Meta and Google. Office landlords have all kinds of tenants — including a growing number of AI start-ups — and they’re waiting to see whether employers in financial and professional services, law and consulting will keep hiring people to oversee artificial intelligence capacity — or whether human resources will dwindle as smarter technology takes over.
Meanwhile, construction spending on offices is down from a 2020 peak, as companies absorb existing stock. But outlays on data centers have increased, according to data from the U.S. Census Bureau analyzed by investor and researcher Paul Kedrosky, and are now level with general office spending.
AI chatbots began to win over ordinary workers, and those workers’ CEOs, in the summer of 2024. Then, this February, advances in the models got so good and enterprise adoption so wide that the scenario spooked some investors, who began selling shares in companies whose business models they figured wouldn’t survive a quick turn to AI. That included some real estate companies — as well as others whose work happens in physical offices.
No
Sarah Mancuso
President, West Region, United States Leasing Advisory, JLL
The narrative is hard to ignore. AI is coming for white collar work, and when those workers go, so does demand for office space. It’s a compelling story. It’s also missing half the picture.
Commercial real estate has lived through disruption before. Email didn’t displace office workers. Mobile technology didn’t either. Each wave transformed how work gets done and, in most cases, the professionals who adapted early captured more market share, not less. AI is the next chapter in that same story.
The displacement thesis assumes that fewer office workers means less office demand. But that conflates two very different things: headcount and economic activity. The companies driving AI’s workforce transformation — the hyperscalers, the frontier AI labs, the infrastructure builders powering this shift — are occupying real estate at a historic pace. The Bay Area alone holds approximately 17 million square feet of AI-occupied space today. JLL Research projects that AI firms could add 20 to 35 million additional square feet by 2030. The sector cast as the victim is actually a primary beneficiary of the very companies accelerating this change.
Inside our own business, AI is doing exactly what every productivity tool before it has done: handling data-heavy work. It is tracking comps, synthesizing market trends, streamlining lease administration and more so that our professionals can focus on what AI cannot replicate. Clients tell us consistently that the primary reason they choose JLL over a competitor is the relationship. The judgment, accountability and trust built over years of navigating complex, bespoke transactions, that is not automatable. It is, however, scalable. And AI is what makes the scaling possible.
Yes, AI will displace some workflows. Morgan Stanley estimates it could generate $34 billion in operating efficiencies across the industry by 2030. But efficiency gains don’t dictate the fate of people who generate them. History shows they tend to unlock new categories of work we haven’t yet named. New advisory services, data center siting, specialized portfolio strategy for AI-native companies: These are already emerging.
The actual risk to commercial real estate is from complacency, not AI. The professionals who will lose ground are those who treat this as a threat rather than the most powerful tool our industry has ever had. Those who move first will define what the next version of this business looks like. And they’ll do it from an office.
Yes
Vikram Malhotra
Managing Director, REITs Analyst at Mizuho Americas
In real estate, everyone focuses on location, location, location. But when it comes to AI, what’s more important than location is jobs, jobs, jobs.
The debate over AI’s impact splits people into two camps: those who imagine Armageddon — AI will eliminate everyone’s job — and the bulls, who say that AI will create a tremendous number of new jobs. The reality is more nuanced, and we are in stage one of figuring out what AI tools can do in terms of efficiency.
To understand the risk, we categorized over 300 occupations by AI risk: high-, medium- and low-risk buckets. High-risk categories include software programmers, certain finance roles, accounting, legal research — all work that can be done either offshore or via technology. If the high-risk bucket dips dramatically, even if we assume the low- and the medium-risk segments continue to grow at typical rates, then “office-using” job growth is going to slow from the 2 to 3 percent per year we’ve seen historically to more like 0.5 percent. A persistent slowing of hiring over the next three to five years has negative implications for multiple real estate sectors, including office, retail and residential.
AI is the excuse for what I’ve been loosely terming as making corporate America efficient again. Whether it’s AI or rationalizing from a period of over-hiring, every corporation will pause and say, “Let’s figure this out.” Just like during Covid, when we were all working remotely, corporates said, “Let me figure out my space needs.” That doesn’t necessarily mean they pause on immediate expansion plans that have been in the works the last six to nine months.
Covid was temporary, but AI could have more long-lasting impacts.
What are we watching? Any potential increase in sublease rates and overall lease renewal rates. There could also be a tick up in sublet volumes as companies stop backfilling roles. If free rents and tenant improvement allowances stay elevated, that tells you demand may be softening even where vacancy looks tight.
Contrary to the popular thesis, it may be coastal markets like New York, Boston and San Francisco that are most exposed, not the Sun Belt. Legal, finance, accounting and consulting firms on the coasts are precisely the organizations that will adopt AI quickest and ask: Do we really need five more younger folks doing research for us, or can we make do with two?
Unfortunately, this is one of those debates where you’re guilty until proven innocent. But until someone announces that the hottest new job is prompt engineer and they need 10,000 of them, it’s going to be hard to prove AI won’t impact job growth.