Digital Realty Trust (DLR) concentration risks

The company's concentration profile is defined by a tiered customer exposure that descends from a high-share group concentration to moderate and low shares at the individual customer level. The twenty largest customers in the portfolio represented approximately 51% of total annualized recurring revenue — a high share by disclosed size and a dependency character, reflecting the degree to which a relatively small number of relationships drive the majority of contracted cash flow. The dependency is meaningful because a tenant departure, contract non-renewal, or capacity reduction decision at any of these twenty accounts would register visibly in the recurring revenue base. Within that group, the top three customers represented approximately 26% of total annualized recurring revenue — a moderate share by disclosed size. This sub-concentration means that even among the top-twenty cohort, the three largest names carry disproportionate weight, making their renewal economics and expansion plans the highest-priority watchpoint in the customer base. At the individual level, the largest single customer accounted for approximately 11.7% of aggregate annualized recurring revenue as of December 31, 2025 — a low share by disclosed size, though it is the highest disclosed individual customer concentration. The declining concentration from group to top-three to largest-single follows a reasonable gradient, suggesting the customer base is not top-heavy to an extreme degree. On balance, the profile describes a business where a manageable group of large technology and enterprise tenants generates the majority of revenue. The primary watchpoints are the renewal pipeline and expansion intentions of the top-twenty customers, particularly the top three.

For the engine’s reasoning on DLR’s current verdict — including which dimensions drove the score — see the per-dimension breakdown.

Concentration counts reflect items disclosed in each peer’s most recent 10-K; disclosed-size classification uses TrendMatrix’s internal 10-K extraction taxonomy.