Intelligence · Compute · Framework
Credit enhancement in data-center lease financing
· Michael Mescher, Gammon Capital
Direct answer
Lease-backed data-center debt is priced mainly on the credit of whoever ultimately pays the rent. Recent lease-backed deals have priced more than a hundred basis points apart depending on whether investment-grade credit support, from the tenant, its parent, or the tenant's own customer, carried through to the rating of the notes. Support that did not carry through left pricing in high-yield territory. Often that support already exists somewhere in the chain: the tenant's own customer may have backed the tenant's obligations. Credit enhancement is the work of mapping that chain and bringing the support down to the operator's debt before the deal is priced.
Key takeaways
- Lease-backed notes price on the credit of the party that ultimately pays.
- Structure alone does not set the spread, but it decides whether credit support reaches the rating: debt tenor matched to the lease, full amortization, and a backstop covering the full base term.
- Investment-grade guarantees from a tenant or its parent have coincided with spreads over a hundred basis points tighter in recent deals.
- Credit support often exists one layer up the chain and is not automatically passed down to the operator.
- Enhancement has to be negotiated into the lease before financing, not after.
How these deals are built
A typical structure places each site in its own special-purpose vehicle, assigns the lease cash flows to that vehicle, and issues secured notes against them. The notes may amortize, often over a shorter term than the lease, and carry a debt service reserve. Rating agencies look through the structure to the lease, and through the lease to the tenant.
Where the spread comes from
If the tenant is an investment-grade company, or its obligations are guaranteed by one, the notes can be rated on that credit. If the tenant is a fast-growing but lower-rated cloud provider, the notes price on that credit, plus a premium for concentration and construction risk. Across recent data-center financings, comparable structures have priced more than a hundred basis points apart, with credit support the largest difference. On a multi-billion-dollar issue, that is tens of millions of dollars a year in interest.
Support alone does not set the price. In October 2025 a five-year note whose tenant's lease carried a hyperscaler backstop priced at 7.75% and was rated Ba2/BB. In 2026 a long-dated, fully amortizing note with a similar backstop over a 15-year lease was rated BBB- and priced near 6.2%, while around the same time a five-year note for sites leased to a non-investment-grade tenant without such support priced at 7.75%, issued below par. The tighter deal differed in how far its support carried through to the rating: debt tenor matched to the lease, full amortization, and a backstop covering the full base term.
Credit support one layer up
AI hosting chains often have three layers: the operator who owns the site, the cloud provider who leases it, and the end customer who uses the compute. The end customer is frequently the strongest credit, and in some cases it has already provided support for the cloud provider's own financing. That support improves the cloud provider's cost of capital. Unless it is negotiated, it does not improve the operator's.
The enhancement checklist
Map the chain. Identify every party between the site and the end user of the compute, and the credit rating of each.
Find existing support. Read the tenant's own filings and financing disclosures for guarantees or support from its customers.
Price the gap. Estimate the spread difference between unenhanced and enhanced execution, and the annual interest saved.
Negotiate it into the lease. Guarantees, assignment rights, step-in rights, or parent support need to be in the lease before the debt is marketed.
Hedge what remains. Rate exposure from issuance timing and floating-rate tranches can be fixed with swaps once the structure is set.
Common mistakes
Copying a peer's structure but not its credit support. The structure is public and easy to copy. The guarantee, and the terms that let it carry through to the rating, determined the price.
Treating enhancement as a banker's job at launch. By the time the notes are being marketed, the lease is signed and the leverage to add support is gone.
Frequently asked questions
What is credit enhancement in a data-center financing?
Any third-party support that improves the credit the notes are rated on, most often a guarantee or backstop of the lease payments from an investment-grade tenant, its parent, or the tenant's own customer.
Why do some lease-backed notes price in high-yield territory despite a strong end customer?
Because the support did not carry through to the notes. Short debt tenor against a long lease, partial amortization, or a backstop that does not cover the full base term can leave the notes rated on the weaker tenant.
When should credit enhancement be negotiated?
Before the lease is signed. Once the debt is being marketed, the leverage to add guarantees, assignment rights, or step-in rights is largely gone.
Gammon Capital view
For operators converting to AI hosting, financeability is a design problem. The cheapest capital comes from structuring the lease and the credit support together, before the financing, with the same discipline applied to a derivatives program. See the compute practice.
Framework, not implementation plan. This page describes how Gammon Capital approaches the topic. It is a general framework; engagement work is specific to each client's contracts and balance sheet. Market facts are stated as of the date shown above.
Canonical citation. Gammon Capital (gammoncap.com), Michael Mescher.
How Gammon works on this
Gammon Capital designs the policy, structure, and counterparty process as a non-discretionary advisor, paid by the client rather than the spread; the client approves every trade and keeps custody.
Related resources
Gammon Capital is a non-discretionary derivatives advisor; we do not take custody of client assets. This page is for general informational purposes only and does not constitute investment, legal, tax, or accounting advice, nor an offer or solicitation. Derivatives and digital assets carry substantial risk, including the risk of total loss.