New quarterly report expected Apr 23, 2026. Report will refresh after results.
D
Digital Realty Trust, Inc.
DLR · Real Estate
Q4 FY2025
Earning ended Dec 31, 2025Report generated Mar 15, 2026
Irreplaceable Infrastructure, Replaceable Margins, Unreplaceable Price
The market has correctly identified the AI tailwind but has completely missed the structural contradiction at the center of this business: the fastest-growing, highest-profile demand — hyperscale AI training — is precisely where Digital Realty has the least pricing power and faces the most credible build-versus-buy threat from its own tenants, while the genuinely sticky, moat-protected business is the unglamorous enterprise colocation that generates no headlines and no analyst enthusiasm.
Flowly Score
Apr 16 Close
-22%$199.05
Fair Value
$155.00
38
Below Averageat $199.05
Business Strength6/10
The physical moat — power access, switching costs, and interconnection density — is genuinely durable, but ROIC persistently below the cost of capital and governance opacity around non-GAAP metrics prevent a higher score; you're owning a real asset with a real moat that earns mediocre returns on incremental capital.
Financial Resilience3/10
An Altman Z below the distress threshold, four years of deeply negative free cash flow, and a debt load that just jumped over a third in a single year paint the picture of a business that cannot fund itself — it is perpetually dependent on the kindness of capital markets, which is fine until it isn't.
Growth & Trajectory7/10
Record bookings two consecutive years, a backlog at full occupancy, and an AI demand wave that is pulling forward years of absorption create genuine revenue visibility — the trajectory is clearly upward, and the shift to power-based reporting signals management is finally running the business the way the economics actually work.
Valuation2/10at $179.61
Even the optimistic DCF scenario doesn't clear the current price, normalized FCF is a fraction of the headline number once the anomalous CapEx pause corrects, and you're paying a premium multiple for a business that has compounded at roughly half its cost of capital for five years — the AI narrative is real, but the price already assumes flawless execution of it.
Risk4/10
The hyperscaler bypass risk isn't theoretical — two of DLR's largest customers have explicitly signaled owned-capacity ambitions on recent earnings calls — and layering on governance red flags, a $24B debt load, multi-jurisdictional regulatory exposure, and legacy infrastructure not architected for liquid-cooled AI racks makes this a business carrying more specific, concrete threats than its occupancy rates imply.
Digital Realty owns something genuinely scarce — power agreements, fiber density, and enterprise interconnection ecosystems in the markets that matter — and that scarcity is becoming more acute, not less, as AI workloads push power requirements to levels that make new site permitting nearly impossible in Tier-1 markets. The business earns real cash, the backlog provides genuine forward visibility, and PlatformDIGITAL's interconnection fabric creates the kind of sticky, compounding customer dependency that degrades slowly even if new leasing slows. The problem is that none of this justifies the current price when ROIC has sat at roughly half the cost of capital for the better part of five years — growing a business that destroys value on each incremental dollar deployed makes the asset base larger without making equity holders richer, and the DCF math reflects that arithmetic brutally. The business is heading toward a bifurcated future. On one side, enterprise and mid-market colocation — the interconnected, multi-workload customers where switching costs are brutal and no hyperscaler wants to compete for a 200kW cage — will remain a durable, cash-generative franchise that earns its keep. On the other side, the hyperscale wholesale business that has driven the last two years of record bookings is inherently transactional, cyclical, and subject to the build-versus-buy calculus of three or four customers who each have more capital than most sovereign nations. The Private AI Exchange Platform and the closed-end fund structure suggest management sees this bifurcation coming and is trying to diversify the business model before the hyperscaler tide turns. The single biggest risk is hyperscaler concentration, and it is more acute than it appears. A handful of cloud giants represent an outsized share of new bookings, and if even one pivots meaningfully toward owned capacity — as signaled publicly by at least two of them — the pre-leasing pipeline that justifies the growth premium dries up while the $24B debt load demands servicing regardless. That isn't a slow-moving risk; it could reprice the stock in a single earnings cycle.