Data Centers Are the Hot New REITs. Should You Invest $1,000?

The market is likely overestimating how much immediate alpha this creates for the sponsor and underestimating how quickly it tightens the competitive moat around the pure-play data center landlords. In the near term, the real beneficiaries are EQIX and DLR because they already own the scarce assets that matter most in AI: power-adjacent, interconnect-rich, pre-leased capacity. A JV into stabilized assets is more of a capital-allocation signal than a growth inflection for O; the financial impact should be modest unless management repeatedly redirects balance-sheet capacity toward this sleeve.

Second-order, the move reinforces a split in the REIT complex between asset classes with pricing power and those still hostage to rates. Data center cash flows are long-duration, so they get bid on growth scarcity, but they also remain highly sensitive to any backup in real yields; a 50-75 bp move in the 10-year can compress multiples faster than NOI growth can offset over a 1-3 month window. Over 6-18 months, the bigger question is whether developers and adjacent power/land owners flood the market with new supply, which would dilute rent growth and reduce the scarcity premium.

Contrarian view: the consensus is treating AI infrastructure as if all exposed real estate should rerate, when in practice the economics accrue unevenly. Stable, investment-grade, triple-net deals are lower risk but also lower upside than the development/platform businesses investors actually want to own. The trade is likely better expressed as relative value inside REITs than as a broad beta bet: if the story is right, DLR/EQIX should outperform O materially; if rates rise or cap rates widen, the whole group gives back quickly.