How Bad Would an AI Bust Be for New York’s Office Market?

In a May 2026 report titled “AI and New York City’s Fiscal Future: Planning Scenarios for an Uncertain Economy,” New York City Comptroller Mark Levine stated that “there is no city in America — and perhaps none on Earth — more exposed to both the promise and peril of artificial intelligence than New York City.”

Noting that “our tax base depends heavily on a financial industry increasingly intertwined with AI,” an analysis by the comptroller’s office laid out five potential scenarios for the future of AI’s effect on the city’s economy, assigning a likelihood percentage to each.

Conclusions included a 35 percent chance that “AI boosts productivity with limited economic disruption,” followed by a 25 percent likelihood of a scenario that includes AI investment fizzling and markets retreating, a 20 percent chance of a “widespread replacement of jobs,” and a 5 percent chance of “labor markets fail[ing] to absorb AI-induced displacement smoothly, with job losses concentrated among white-collar jobs.”

The possibility that “AI drives faster growth, wages and prosperity,” with no negative effects cited, was assigned just 15 percent.

All could be forgiven their most optimistic or pessimistic assumptions about the effects of the technology’s growth on both the economy and humanity at large given the flurry of progress, activity and controversy surrounding AI. That includes Goldman Sachs projecting $1 trillion in global AI-related investment in 2026, including $581 billion in the U.S., while also noting that AI has eliminated 16,000 net jobs every month over the past year; the International Monetary Fund saying just this month that an AI-related market correction, a stark increase in government borrowing costs and the threat of cyberattacks are “plausible near-term catalysts for a sharp tightening of financial conditions and market turmoil”; and senior Anthropic researcher Evan Hubinger claiming a greater than 10 percent chance that AI could “kill all humans” over the next decade.

But, if there are potential AI dangers barreling down on us from all sides, New York City’s office sector has so far been sheltered from the target the technology is supposedly painting on our backs.

According to Savills, AI companies leased 1.7 million square feet of office space in Manhattan in the first half of 2026, accounting for 8.2 percent of all new leasing in the borough — up from 0.7 percent in all of 2020 — and eclipsing total new leasing by AI companies for all of 2024 and 2025 combined.

The company also noted that AI leasing in the borough has shifted from subleasing to direct deals, with the share of AI subleases falling from 42.3 percent to 17.5 percent when comparing cumulative totals from 2020 through 2024 against the period including 2025 and the first half of 2026.

The average AI lease term length has also skyrocketed in Manhattan, from 49 months in 2020 to 90.9 months in the first half of 2026.

Given all these conflicting signals, those focused on the health and fate of New York’s office market seem to be mostly filtering out the fears and noise to the best of their abilities, lasering in instead on the sector’s current prosperity and the role of AI within it.

“The level of excitement in the market right now for AI as part of the leasing landscape of Manhattan is justifiably high,” said Gabe Marans, a vice chairman at Savills. “New York more so than any other market, with the possible exception of the Bay Area, is the capital of applied AI — AI businesses that have an often industry-specific application.”

But, even with the massive relative increase in AI-occupied office space and AI’s prominence in the global conversation overall, it’s important to keep it in perspective. One figure from Savills in particular puts AI-related fears for the office sector closer to the back burner.

The company’s research department estimates that even with the notable increases in AI leasing in 2026, those companies currently occupy just around 1 percent of Manhattan office space. That makes any fears around an AI crash a far more manageable threat for the health of the sector in New York City.

This allows the city’s office market veterans the luxury of reveling in AI’s contributions to sector growth without having to fret too much about the implications of a potential implosion in the sector.

While Marans acknowledges concerns, especially since AI has seemingly evolved from “zero to 60 for an industry that did not exist four years ago,” the relatively small portion of the office sector AI currently occupies provides breathing room even in the event of a worst-case scenario.

“Even if a bubble was starting to percolate, as a percentage of total leasing in New York, it is still rather muted,” said Marans. “If we woke up tomorrow to news that this AI bubble burst, New York City’s office market would barely skip a beat and would continue marching forward.”

That said, given the rate of AI’s growth in the sector, that 1 percent figure will likely increase soon. And, while the percentage number may currently be small, the square footage the space occupies — around 4.2 million square feet and growing, according to Savills — means a disruption would have oversize effects on those whose properties are more flush with AI tenants.

“The biggest risk right now for Class B office owners, myself included, is that we’re leasing space to AI tenants,” said Jeff Gural, chairman and principal of GFP Real Estate. “When we had the dot-com boom, I didn’t lease space to any of them, so I didn’t get hurt. But, today, you have no choice but to lease space to AI tenants, because they’re the ones taking most of the space.”

Gural, like most, acknowledges that the reality of a potential AI shakeout would likely be that as with any new and evolving industry — especially one that is growing this quickly — there will be winners and losers.

Whether there would be a real bust would be determined by how many winners and losers there are, and their size. These mark some of the most essential and impactful unknowns for the office industry in the coming years.

“When I ask the experts, they all say the same thing: Some AI companies will make it, and some of them won’t. So it’s just a guess,” said Gural. “Most of them have enough money to last three years. They’ll give you a letter of credit for a year’s rent. We’ll see how that plays out.”

Josh Winefsky, a partner at Herbert Smith Freehills Kramer, pointed out that even if an AI shakeout resulted in fewer AI companies overall, that might not lead to a reduction in office demand.

“If 1,000 companies took a million feet of space, but in three years those 1,000 companies were whittled down to 400, those 400 could still need that million feet, because the ones that have survived will have figured it out,” said Winefsky. “They will have evolved and implemented the internal protocols that would hopefully satisfy the security and existential threats everybody’s concerned about.”

Others hold a similar perspective that, while acknowledging a range of possibilities, focuses more on how AI is already so ingrained in the corporate world and beyond that it’s clearly here to stay. So the best we can all do is adapt and prepare for any eventuality.

Steve Durels, executive vice president and director of leasing and real property for SL Green, the city’s largest office landlord, said that the “genie is out of the bottle” with AI. As such, ensuring that the technology is properly assessed, managed and regulated is key.

Durels also believes some of the general AI fears relate differently to New York City.

“The fear people have of AI eliminating jobs is not a real fear for New York City because these are not call center service jobs that dominate our marketplace. We’re the front office,” said Durels. “People in New York City can’t be replicated or replaced by AI.”

Given that, Durels takes an optimistic view overall of AI’s effect on the city’s economy.

“More businesses will be created as a result of AI, and people will become more productive as a result of AI,” said Durels. “To the extent that jobs are changed or eliminated, they’re more the back office jobs that are housed outside of Manhattan.”

Many have echoed this confidence, including those with clear memories of the dot-com bubble, when the enthusiasm brought about by exciting new technology and its long-term economic potential were quickly brought to a dispiriting reality.

The rapid rise of AI can also make comparisons to certain recent financial collapses seem too tightly aligned for comfort.

But Rich Monopoli, senior vice president at owner BXP, points out that while there are some frighteningly close similarities, there are notable differences as well.

“I lived through the late `90s and the dot-com era, as did BXP. I would make one comparison, and that’s the health of our occupancy,” said Monopoli. “In the 1990s, two guys with an internet business plan could raise $100 million, then sign a lease for 50,000 square feet, thinking they were going to have 500 people in six months. There was a substantive grab of space relative to the bodies, and it was unjustified. When it hit the fan, that was reflected in the subleasing numbers.”

The AI industry as we’ve seen it so far has been smarter and more conscientious regarding office space, Monopoli said.

“AI companies have the bodies to fill the space,” said Monopoli. “It’s a healthier occupancy relative to what we saw back then.”

And Michael Cohen, a principal at Williams Equities, another owner, noted that the AI industry isn’t the only one making smarter choices.

“A lot of these dot-com companies didn’t have a tangible product, revenues or clientele,” said Cohen. “With the experience of the dot-com years, many of us are weighing the decisions we make carefully when leasing to AI companies, reviewing their balance sheets and their [profit and loss], as we determine how much risk to take on a particular tenant.”

With all this taken into account, New York’s office community feels strongly that even as AI sends us into a future where literally anything could happen, the city’s office revival should continue mostly unabated.

“In the event that AI is on its way to a bust, what kind of impact will it have on the market?” Savills’ Marans said. “While the growth of AI has been super fast, it’s not going to move the needle in a significant way one way or another.”

Larry Getlen can be reached at lgetlen@commercialobserver.com.