AI Is Booming in Commercial Real Estate

Manhattan is having its best leasing year since 2000. San Francisco's office market — written off two years ago — is quietly stabilizing. Across the country, industrial real estate is seeing demand that nobody projected five years ago. The common thread in all of it is AI.

And now, 70 cities want to pump the brakes.

The Boom Is Real

AI firms leased 1 million square feet of Manhattan office space in Q1 2026 alone — more than they leased in all of 2025. They accounted for 56% of all tech-sector leasing in the quarter. Total Manhattan leasing activity has hit 19.6 million square feet year-to-date, with supply at its lowest level since October 2020 and rents beginning to climb.

San Francisco tells a similar story. More than 90 AI companies are now based in the Bay Area, occupying 13.4% of the city's leased office space, according to Avison Young. For a market that spent three years on commercial real estate's version of a do-not-resuscitate list, that's a meaningful reversal.

Beyond office space, AI is reshaping where industrial real estate gets built. Massive data centers are going up in Abilene, Texas; Mount Pleasant, Wisconsin; Holly Ridge, Louisiana — places that wouldn't have appeared on a commercial real estate map five years ago. The infrastructure buildout required to run AI at scale is creating demand in markets that have never seen it.

The Resistance Is Also Real

More than 70 cities and counties across the U.S. have now announced temporary or permanent bans on new data center construction. The most recent is Seattle, where the city council unanimously approved a one-year moratorium on June 9 after a group of companies proposed five large data centers with a combined peak power demand of 369 megawatts — roughly one-third of the city's average daily electricity consumption.

Seattle isn't an outlier. Illinois Governor JB Pritzker suspended new state tax incentives for data centers this month. The New York State Legislature passed a bill calling for a statewide one-year moratorium on data centers with peak energy demand above 20 megawatts — the first such bill in the country. It's currently on Governor Hochul's desk.

The objections aren't ideological. They're practical. Data centers consume enormous amounts of power and water. When a single proposed project represents a third of a city's daily electricity use, local officials have legitimate infrastructure questions that don't have fast answers.

What This Means for the Broader AI Buildout

The tension here isn't going away. AI's infrastructure requirements are real and growing, and the communities being asked to host that infrastructure are starting to push back on the terms. The companies building AI at scale need somewhere to put the compute. The places being asked to absorb that compute are discovering that the economic benefits don't always offset the strain on local power grids, water systems, and existing residents.

For the commercial real estate market, this creates a geography problem. The markets with the most favorable conditions for data center development — available land, power infrastructure, favorable tax policy — are getting more crowded and more contested simultaneously. The markets pushing back are, in some cases, exactly the ones where demand is highest.

The office leasing story and the data center story are connected. You can't run the AI companies filling up Manhattan without the infrastructure those companies depend on. If that infrastructure gets constrained by regulation, the pace of the broader AI expansion slows with it.

The Number Worth Watching

Manhattan's leasing trajectory is genuinely impressive. But the more consequential number right now is 70 — the cities that have moved to restrict data center construction. That figure has been growing steadily, and the legislative momentum in New York and Illinois suggests it will keep growing.

The AI real estate boom is real. So is the friction building around it. The firms and investors paying attention to both sides of that equation are going to make better decisions than the ones focused only on the headline numbers.