Two years ago, the biggest question for office landlords was how much space companies would need post-pandemic. Now that question has been replaced by a harder one: what happens to office demand as AI changes the workforce itself? Speaking at industry events this week, including Bisnow’s 2026 New York Commercial Leasing and Asset Management Conference, real estate executives said that uncertainty is now the dominant force behind lease negotiations across Manhattan.
Even companies with long, stable histories say it’s gotten difficult to plan ahead. KPMG holds a secure long-term lease in Manhattan but has decisions looming on offices in other cities within five years, and the firm’s real estate leadership says forecasting space needs even a few years out, let alone by 2031, has become far harder with AI’s labor impact still unclear, according to Bisnow.
That’s not an isolated concern. A recent survey of senior executives found 60% say AI has made it harder to predict their office space needs over the next two years, and 88% say flexibility in real estate has become more important as a result.
A Tighter Market Shifts the Power Balance
That uncertainty is colliding with a market that’s tightening fast. Manhattan is headed for its best leasing year since 2000, and available space has dropped to its lowest level since 2020, sitting at 12.5% borough-wide and just 3.5% in top-tier buildings, according to Colliers. Landlords say large spaces are now fielding multiple competing offers, a dramatic change from the tenant-friendly conditions of a few years ago.
With that leverage, ownership groups are pulling back the flexible terms, like contraction and early termination options, that became common after 2020. Firms including Tishman Speyer and Silverstein Properties say tenants are now expected to commit to space that accounts for future growth upfront rather than negotiating room to scale down later. Free-rent periods and tenant improvement allowances have also shrunk, with one major landlord, SL Green, reporting drops from 18 months of free rent to roughly 14–16, and improvement allowances falling from the $150–$165 per-square-foot range to about $145–$150.
Tech’s Growth Comes With Landlord Caution
AI isn’t just complicating forecasts for existing tenants — it’s also changing who gets approved for new leases. Technology firms are now the second-most active leasing sector after financial services, with a large share of that activity driven by AI startups. But unlike legal or financial tenants with decades of operating history, many AI companies are new enough that landlords say they have to dig deeper into leadership and financial backing before signing off, and some deals have fallen apart entirely over those concerns.
One real estate attorney at the Bisnow conference likened the current pace of AI-driven leasing to the dot-com boom of the early 2000s, a comparison that underscores why landlords are treating this wave of tenants more cautiously than past growth cycles.
What This Means Going Forward
Flexibility hasn’t vanished entirely. Landlords say they’ll still extend more favorable terms to tenants with strong credit and long track records, particularly on larger transactions, but typically only in exchange for firmer long-term commitments. For most companies, though, the message from this week’s industry gatherings was consistent: with space tightening and AI adding a new layer of unpredictability, landlords have little incentive to offer the kind of flexibility that defined the post-pandemic years.
