Signing a ten-year office lease has always required a leap of faith about the future. Now, artificial intelligence is making that leap considerably harder to take. New survey data suggests that AI isn’t just changing how people work inside an office; it’s changing whether companies dare to commit to one at all, and flexible workspace is emerging as the practical way around the problem.
The two-year crystal ball problem
A recent IWG-commissioned survey of 1,000 senior executives found that 60% said AI has made it harder to predict how much office space their organization will need over the next two years. An even larger share, 88%, said flexibility in their workspace or real estate strategy has become more important as a result.
That’s a striking admission from the people who actually sign the leases. Traditional office planning has always relied on a fairly simple formula: expected headcount, a desk ratio, and a workplace strategy. AI complicates every part of that equation. A company rolling out AI tools might not know whether the technology will shrink certain teams, create demand for new specialists in areas like data governance or AI oversight, or simply fail to deliver the productivity gains it promised. Unlike past disruptions, which tended to hit specific sectors, AI’s effects are proving to vary wildly even between departments in the same company: fewer people needed in document review or first-draft research, potentially more in compliance, client assurance, and technical oversight.
That unpredictability collides head-on with how conventional office leases work. A typical lease can run for a decade or more, often with significant fit-out costs and an implicit bet that the tenant’s size and shape will stay roughly constant. Break clauses exist, but they’re often rare and expensive to use. When a management team can’t confidently forecast headcount two years out, promising to fill a fixed footprint for ten becomes a much riskier commitment.
Betting small instead of betting the building
This is where flexible workspace changes the calculation. Rather than choosing between “take the office” or “don’t,” companies can hold a smaller core headquarters and use flexible space, whether that’s coworking memberships, managed offices, or serviced suites, to absorb whatever comes next: a growth spurt, a shrinking back office, a new market entry, or a project team that only needs to exist for six months.
The economics support the shift. A company that overcommits to conventional space ends up paying for empty desks for years. One that undercommits risks an expensive, disruptive relocation. Flexible space offers a partial hedge against both outcomes, and increasingly, boards seem to be treating that hedge as worth paying for, even where the per-desk cost of serviced space runs higher than a well-negotiated conventional lease.
Follow the money: What IWG and Orega are seeing
This isn’t just a theory; the commercial data backs it up. IWG, which owns brands including Regus, Spaces, and HQ, reported a record 782 location openings in its 2025 annual results. Its most recent interim results, published in August 2026, showed group revenue up 4% year-on-year to $958 million, with system-wide revenue up 9% to $1.166 billion. Notably, the group’s managed-and-franchised business, which generates fee income without IWG carrying the full leasehold risk, grew fee income by 70% in the first quarter, a sign that landlords themselves are increasingly betting on the flex model.
UK-focused operator Orega tells a similar story. Its 2026 Flex Report found that 81% of office landlords, advisers, and asset managers plan further flex-space expansion within the next one to three years, and 36% named economic uncertainty as the single biggest driver of flex growth over the next five years. That’s a meaningful shift: flexible space is no longer viewed as a niche product for freelancers and early-stage startups. Landlords themselves increasingly see it as a way to keep conventional buildings relevant to occupiers who no longer want to make ten-year bets.
The lease isn’t dead, it just got a roommate
None of this means the long lease is disappearing. Large, stable occupiers with specialist requirements or an appetite for bespoke headquarters will keep signing conventional space, particularly in prime buildings. But more companies appear to be splitting their portfolios: a core estate for the functions that genuinely need permanence, paired with managed offices, coworking memberships, and on-demand meeting space for everything else, sometimes called the corporate “donut” model.
The bigger picture is that AI isn’t necessarily shrinking the office market. It’s making demand harder to forecast, which is a very different problem, and one that rewards operators and landlords who can offer optionality rather than long-term certainty.
The takeaway
As AI reshapes how, and how much, companies staff their teams, the people making real estate decisions are quietly concluding that flexibility isn’t just a nice-to-have anymore. It’s a way to avoid making an expensive bet on a future nobody can currently see clearly. For landlords and flex operators, that uncertainty is turning into one of the more durable sources of demand in today’s office market.
