Digital Realty (DLR) Q1 2026 Earnings Transcript

Digital Realty delivered a strong quarter with core FFO of $2.04 per share, up 15% year over year, while raising full-year 2026 guidance by $0.10 to $8.10 per share. Leasing was exceptionally strong at $707 million annualized, including a company-record 200MW AI inference lease in Charlotte, pushing backlog to a record $1.8 billion and development under construction to 1.2GW. The company also lowered leverage to 4.7x, expanded its pipeline to $16.5 billion, and increased 2026 CapEx guidance to $3.5 billion-$4.0 billion.

Analysis

DLR’s print is less about a one-quarter earnings beat than a multi-year capacity lock-up. The key second-order effect is that the company is now converting scarcity in power-constrained metros into longer-duration cash flow with embedded inflation protection: the mix shift toward inference, enterprise interconnect, and 15-year hyperscale contracts makes the revenue stream materially stickier than the market likely assumes. The raised capex and backlog also imply that near-term FFO acceleration is being deliberately traded for a much larger 2027-28 visibility wedge, which should support a higher multiple if execution holds.

The competitive read-through is negative for smaller colo and less-connected regional operators. DLR is proving it can monetize both ends of the demand curve simultaneously — sub-megawatt enterprise and 100MW+ hyperscale — while also using private capital to de-risk growth, which raises the bar for anyone trying to compete purely on balance-sheet capacity. Suppliers of liquid cooling, electrical gear, and mission-critical construction labor remain the bottleneck winners, but they also create the most obvious margin risk if lead times or wages re-accelerate.

The main risk is not demand collapse; it is timing slippage. The market may have to re-rate the stock lower on the next 1-2 quarters if commencements lag bookings again, because the equity story is becoming more explicitly back-end loaded. A second-order bear case is that rising utility/interconnection complexity and community pushback inflate basis faster than market rents, compressing the spread between “headline demand” and actual delivered returns.

Contrarian angle: consensus may still be underestimating how much AI inference looks like enterprise connectivity, not just hyperscale GPU compute. If that’s right, the addressable market widens meaningfully into smaller committed blocks, which should improve utilization of legacy assets and increase renewal pricing power faster than the market models. The stock is likely less a ‘data center REIT’ and more a ‘scarcity infrastructure compounder’ now.

Key Decisions for Investors

• Long DLR versus short a lower-quality colo peer basket over 3-6 months; the pair should benefit as capital-intensive, power-constrained incumbents win the largest AI/inference deployments while weaker names face margin compression and slower lease conversion.
• Buy DLR on any 5-8% pullback after the next quarterly report; use a 6-12 month horizon and target upside from backlog-driven 2027 visibility, with risk defined by commencement delays or a sharp capex reset.
• For event-driven exposure, sell out-of-the-money puts in DLR 2-3 months out if implied vol spikes on concerns about capex or utility timing; the thesis is that demand is real but timing noise creates attractive premium.
• Pair long DLR / short an AI-enabler that is already priced for perfection in infrastructure growth, such as NVDA, if you want to express a relative bet that real estate scarcity monetization is underappreciated versus hardware enthusiasm; keep sizing modest due to factor risk.
• Avoid chasing after a strong print if your mandate is short-duration alpha; the better entry is on any evidence of another commencement delay, which should be transient rather than fundamental and could provide a better risk/reward entry point.