AI driving deep divergence across real estate markets as highest-exposure cities show strongest resilience

This is less a thesis on "AI helps real estate" than a thesis on where economic rent accrues: premium space in AI hubs, not the sector average. The first-order winner is brokerage/advisory firms with exposure to leasing, valuation, and repositioning work; the second-order winner is owners of top-tier office and data-center-adjacent assets that can extract pricing from scarce, power-rich locations. Legacy suburban office, weaker credit tenants, and CMBS tied to commodity buildings are the obvious losers as the market rewards adaptability rather than footprint reduction alone.

For JLL specifically, the upside is real but not linear. If AI-driven occupancy remains concentrated in a handful of markets, JLL should see better leasing and capital-markets activity, but the bigger revenue lever is still transaction velocity, which lags sentiment by 1-3 quarters. The more important read-through is that "AI exposure" can be positive for revenue per square foot in the right submarkets, while simultaneously destroying value in the wrong ones; that bifurcation should widen appraisal gaps and slow refinancing for Class B office over the next 6-18 months.

The main risk is that the market extrapolates a local, coastal-office story into a broad CRE recovery. If AI hiring/funding cools, the demand support becomes a narrow hub effect rather than a structural upcycle, and the enthusiasm fades quickly. The cleaner contrarian is that the trade is not "long office" but "long scarcity": trophy assets, data centers, and the advisors who intermediate migration, versus short obsolete office and lenders exposed to it.