BTIG starts DLR, EQIX at Buy, calls data center demand a multi-decade supercycle

BTIG initiated coverage of Digital Realty Trust and Equinix with Buy ratings, arguing AI-driven demand can support a decades-long data center “supercycle.” The firm set a $215 price target for Digital Realty and $1,210 for Equinix, citing record 2026 hyperscaler capex of $650B+ (+72% YoY) and data center rents up 63% over five years. Despite supply expected to rise 86% through 2030 (global spend ~$5.1T), BTIG expects power constraints and long lead times to limit cyclical overshoot.

Analysis

The cleanest expression is not a generic long on “data centers,” but a barbell between scarcity assets and capital-intensive growth. EQIX should retain a premium because interconnection density and enterprise mix make its pricing power less dependent on any single hyperscaler, while DLR is more exposed to lease timing, development carry, and the risk that the market discounts its growth as closer to infrastructure than software-like compounding.

The bigger second-order winners are not the REITs alone. Power-constrained names with leverage to grid buildout and thermal management — VRT, ETN, PWR, and selected utilities with large interconnection queues — may see the most durable order visibility, because the binding constraint is increasingly transformers, switchgear, and transmission, not land or shell capacity. That shift should show up first in backlog and lead times over the next 2-6 quarters, then in pricing power if supply remains tight.

The main risk is that “demand” headlines mask deteriorating economics: higher rates, expensive build costs, and a long development cycle can compress spreads even if occupancy stays high. Over 6-18 months, watch whether stabilized yields stay above WACC and whether hyperscalers start internalizing more of the stack; if rent growth slows or leasing spreads flatten, the supercycle multiple can de-rate quickly.

Contrarian view: consensus may be overstating how much of AI spend accrues to landlords versus chip vendors and cloud operators. If model efficiency improves or inference shifts toward edge/local deployment, the watts-per-dollar thesis weakens, making today’s supply pipeline look less like a decade-long shortage and more like a late-cycle capital rush.

Key Decisions for Investors

• Long EQIX / short DLR for the next 3-6 months: own the higher-quality moat, short the more development- and financing-sensitive name; target relative outperformance if power scarcity keeps supporting core-market rents.
• Buy VRT or ETN on pullbacks as a picks-and-shovels trade on the power bottleneck; risk/reward is better than the REITs if backlog conversion continues and lead times stay stretched.
• Use XLRE as a hedge against the broader real estate beta while expressing the AI-specific view through EQIX: long EQIX, short XLRE on a 1:1 dollar basis if you want to isolate sector dispersion.
• Set an alert on DLR if development yields or leasing spreads deteriorate in the next two earnings prints; that would falsify the scarcity-premium thesis and argue for covering the relative short.
• If you need options, favor 6-9 month call spreads on EQIX rather than outright calls: the upside is driven by steady multiple support, but the trade can stall if rates back up.