The most important number in the Manhattan office market is not the 719,200 square feet AI firms have leased this year. It is the 12 million square feet of sublease space that has vanished since late 2022.
Sublease inventory fell below 11 million square feet in the second quarter, according to JLL, less than half the 23 million square foot peak. That is not just absorption. It is a structural shift in how landlords and tenants are pricing risk.
AI firms like Tennr, Uber, Datasite, and Bank of Montreal are taking space. But the real driver of the sublease decline is landlords pulling back available space to pursue direct deals at higher rents. Jamie Katcher, JLL executive managing director, put it plainly: landlords are increasingly pulling back sublease space to pursue direct deals at substantially higher rents.
This is not a demand story alone. It is a pricing story.
When a landlord takes back sublease space, they are making a calculated bet that the market will pay more for a direct lease than a sublease. That bet only works if the landlord has enough confidence in tenant demand and enough patience to wait for the right deal. It also works only if the landlord has the balance sheet to carry vacant space while waiting.
The capital markets implication is straightforward. Sublease space is a discount product. It signals that the tenant who originally leased the space overpaid or overcommitted. It also signals that the market clearing price for that space is below the original lease rate. When landlords pull sublease space, they are effectively saying the discount is no longer necessary. The market clearing price has risen.
That is a bullish signal for office valuations, but only for the right buildings. The buildings attracting AI and financial services tenants are the ones with modern infrastructure, good locations, and flexible floor plates. The buildings that still have sublease overhang are likely the ones that lack those attributes.
The bifurcation is intensifying. Capital is not flowing to all office assets. It is flowing to assets where the landlord can command a direct lease premium over sublease. That premium is the difference between a building that can refinance and one that cannot.
For lenders, the sublease decline is a positive data point. Sublease space depresses rents and compresses net operating income. Less sublease inventory means less downward pressure on in-place rents and less risk that a borrower's cash flow falls short of debt service. But lenders should be careful not to extrapolate. The sublease decline is concentrated in the same submarkets that are already seeing the most leasing velocity. It is not a broad market recovery.
For owners with maturities, the sublease decline is a reason to be cautiously optimistic. If the market clearing price for direct space is rising, then the basis at which a building can be refinanced may be higher than it was six months ago. But that only matters if the building is in the right submarket and has the right sponsor.
For tenants, the sublease decline means the era of cheap sublease deals is ending. Tenants who waited for sublease bargains may find that the best deals have already been taken. The window for opportunistic sublease arbitrage is closing.
The next phase of the Manhattan office market will not be defined by how much sublease space is available. It will be defined by which landlords have the capital and conviction to hold out for higher direct rents, and which tenants are willing to pay them.
Liquidity has returned to the office market, but only where the pricing is defensible. The sublease decline is proof that defensible pricing is coming back. The question is how much of it is real demand and how much is landlords simply refusing to sell at a discount.