A short list of investment-grade global cloud platforms and AI compute platforms anchor most hyperscale data center construction in 2026. The roster includes the largest U.S. public cloud platforms, the major AI training platforms, and a handful of large enterprise and social platforms. Concentration in any single tenant is a meaningful underwriting variable.
A concentrated tenant universe
The hyperscale tenant universe is narrow. The largest U.S. public cloud platforms (the three dominant infrastructure-as-a-service providers) account for the bulk of new build-to-suit absorption. The major AI training platforms (a small set of dedicated AI infrastructure operators and the cloud platforms' own AI divisions) account for most of the recent step-change demand. A handful of large enterprise platforms (the largest social, search, video streaming, and SaaS operators) round out the bench.
Almost every name in this universe is investment-grade at the parent level. Lease guaranties, however, are sometimes signed at a subsidiary level rather than the parent, and lenders pay close attention to the legal entity that signs the lease and whether the parent guaranty is express or implied.
Why concentration risk matters
A typical institutional data center sponsor's portfolio carries meaningful concentration in two or three tenants. CTL perm lenders explicitly cap their per-tenant exposure across all CTL deals in their book, which can affect a sponsor's ability to take CTL perm on multiple deals to the same tenant. Sponsors planning multi-deal CTL programs should diversify their lender options accordingly.
Sources
- 2024 United States Data Center Energy Usage Report · Lawrence Berkeley National Laboratory (LBNL) · 2024
- Capital Markets Special Report: Insurance Industry CRE Mortgage Holdings · National Association of Insurance Commissioners (NAIC) · 2025-Q4
- Data Center Energy Use: GAO Reports on Power Demand & Federal Programs · U.S. Government Accountability Office (GAO) · current