S14.3 · Real Estate & Built Environment
Warehouse, logistics and data-center property where AI capex is directly repricing land near power and fiber.
Industrial and logistics real estate is the development and ownership of warehouse, logistics and data-infrastructure property. No single aggregate captures the segment reliably; the working proxies are hyperscaler AI data-center capex of $775-800B for 2026 (AI Capital Advisory) and industrial and warehouse absorption data from CBRE and Cushman & Wakefield showing data centers pulling a second wave of adjacent industrial demand. Nowhere else in the sector does AI reprice real estate this directly: every gigawatt of data-center construction generates roughly 2 million square feet of adjacent industrial demand (Link Logistics, 2026), and the binding constraint on that demand is power availability, not underlying appetite.
After the 2022-2024 oversupply, industrial vacancy stabilized and began tightening through 2025-2026, while data-center absorption compounds at a double-digit annual rate. The two demand pools locate differently — logistics near ports and population centers, meaning the Inland Empire, Dallas, Atlanta and New Jersey; data centers near power and fiber availability, notably Northern Virginia, Phoenix and Columbus. So does ownership: a small number of large developers dominate logistics, while data-center product concentrates among specialist REITs and hyperscalers building for their own account.
Upstream sit power utilities and land; then developer; then hyperscaler or third-party logistics tenant; with mechanical-electrical-plumbing contracting downstream for fit-out and an adjacent interconnection layer tying the property to the grid and to fiber networks. Capital intensity is extreme in the data-center sub-segment, but it buys something logistics does not have: power-infrastructure investment supporting long-lease, recurring net operating income that reads closer to a utility contract than to a conventional lease. Industrial logistics runs shorter, more transactional lease cycles — the traditional warehouse model.
Both adjacencies that matter — mechanical-electrical-plumbing contracting for build-out, and power and energy development for on-site and adjacent generation — point the same way: logistics owners pushing outward to capture data-center-like economics through power co-location. That is new-dollar territory, not margin defense.
The AI money entering this segment is capex, not automation. It reprices land near power and fiber and drags industrial absorption along with it; no labor cost line collapses, because nothing is being automated — the effect is pure demand. What that demand is building is a "power-adjacent industrial" asset class with its own underwriting logic, distinct from traditional logistics: where highway access used to price industrial land, power capacity now does. All of this is current rather than forward-looking, and it is the sector's sharpest case of AI infrastructure spend flowing straight into real asset values rather than into labor displacement.