Underwriting a Gulf Data Centre: Offtake, Tenants and Covenants

Introduction

To an equity investor, a data centre is a building and a business; to a lender, it is a contract. The lender does not share in the upside of rising demand or the asset’s appreciation; it is repaid from the cash flow the data centre generates, and that cash flow comes from the contract under which a tenant pays to use the capacity. The lender’s task in underwriting a data centre is therefore, above all, to assess that contract, the offtake, the creditworthiness of the tenant who signs it, and the lease terms and covenants that determine how secure and durable the payments are. The physical asset matters, but chiefly as the thing that supports and could substitute for the contracted cash flow; the contract is what the lender is actually lending against, and the quality of that contract determines whether, how much, and at what cost the asset can be financed.

This paper sets out the diligence and underwriting model for a Gulf data centre from a lender’s perspective. It is written for the credit and project-finance teams of a lender, a bank or a private-credit fund, considering debt against a Gulf data centre, and for the equity sponsors who must understand what lenders will require. Its purpose is to organise the underwriting around the factors that actually determine financeability: the offtake and tenant covenant first, then the lease terms and covenants, then the debt-service coverage under stress, and then the power and completion risks that can derail an otherwise sound deal. The central message is that data-centre underwriting is contract underwriting, that the tenant covenant and the offtake dominate the analysis, and that the asset, the power and the construction are assessed in service of the question whether the contracted cash flow will be paid and will service the debt.

The companion paper in this cluster makes the strategic case for deploying capital into the Gulf data-centre boom and identifies power and offtake as the dominant value drivers; this paper turns to the lender’s underwriting of a specific asset and develops the offtake and covenant analysis in depth. Where the strategic paper asks where to deploy, this paper asks how to underwrite the debt against a particular data centre, which is the discipline that protects the lender and, in turn, determines the leverage and cost available to the sponsor. The two are complementary views of the same asset, from the strategic and the credit perspectives.

The paper makes three contributions. First, it weights the pillars of financeability and shows that the tenant covenant and offtake dominate. Second, it shows how covenant quality drives the leverage and cost of debt, and identifies the lease terms and covenants that protect the lender. Third, it models debt-service coverage under combined stress and sets out how power and completion risk are underwritten as gating items. Throughout, the figures are modelled and clearly labelled; they illustrate the underwriting mechanics rather than forecasting outcomes, and they are not forecasts.

Results And Discussion

This section presents the framework in the order of the propositions: the pillars of financeability (Proposition 1); the covenant-leverage relationship (Proposition 2); the protective lease terms (Proposition 3); the debt-service coverage under stress (Proposition 4); and the underwriting of power and completion risk (Proposition 5).

This contract-centric framing also has a reassuring corollary for the lender new to data centres: much of the underwriting is familiar. A lender that has financed other contracted-cash-flow assets, let real estate, infrastructure with offtake agreements, leased equipment, already possesses the core skills the framework requires, assessing a counterparty’s credit, scrutinising a long-term contract, testing coverage, structuring covenants and reserves. The data-centre-specific elements, the power intensity, the technology obsolescence, the specialised re-letting market, are additions to this familiar core rather than a wholly new discipline. This means a capable contracted-cash-flow lender can extend into data centres by adding the sector-specific diligence to its existing underwriting skills, rather than building an entirely new capability, provided it engages the technical and power expertise the asset requires. Recognising the familiar core within the novel asset helps a lender approach data centres with appropriate confidence, neither overawed by the technology nor blind to the genuinely new risks, which is the balanced posture the framework supports.

4.0a Why a Data Centre Is, to a Lender, a Contract

Before weighting the pillars, it is worth establishing the proposition that organises them: that a data centre, to a lender, is fundamentally a contract rather than a building. The building has value, but its value to the lender derives almost entirely from the contracted cash flow it houses, because the lender is repaid from that cash flow and only secondarily from the asset if the cash flow fails. A data centre with no tenant is, to a lender, a large, specialised, power-hungry structure of uncertain value and limited alternative use; the same building leased to a strong tenant on a long lease is a secure, bond-like cash flow worth lending against substantially. The transformation from the first to the second is effected entirely by the contract, which is why the contract, not the building, is the object of the lender’s underwriting. This is the sense in which data-centre lending is contract underwriting, and it is the proposition from which the pillar weighting and the whole framework follow.

This contract-centric view also explains why the lender’s diligence differs from the equity investor’s in emphasis. The equity investor must form a view on the asset’s long-term value, its growth, its competitive position, its eventual sale, because it owns those outcomes; the lender must form a view chiefly on the security and durability of the contracted payments over the life of its loan, because that is what repays it. Both examine the asset, the tenant, the power and the market, but they weight them differently: the equity toward value and growth, the lender toward the security of the contracted cash flow. Recognising this difference in emphasis is the first step in adopting the lender’s perspective, and it is why a lender cannot simply borrow the equity sponsor’s analysis but must conduct its own, weighted toward the downside the sponsor is less focused on.

The Pillars of Financeability

Underwriting begins by identifying what makes a data centre financeable. Figure 1 weights the pillars.

Figure 1. What Makes a Gulf Data Centre Financeable

Indicative; the tenant covenant and lease dominate, ahead of power, construction and sponsor quality.

The figure supports Proposition 1. The largest weight in underwriting a data centre falls on the tenant covenant, the creditworthiness of the entity that has contracted to pay for the capacity, followed by the lease term and terms, then power security, then construction and completion, then the sponsor and operator. The ordering reflects the lender’s fundamental position: it is repaid from the contracted cash flow, so the security and durability of that cash flow, set by the tenant’s creditworthiness and the lease, weigh most heavily. The physical asset, the power and the construction matter, but they matter as the things that enable and could substitute for the contracted payments, not as the primary object of the lending. The practical lesson is that a lender should concentrate its underwriting effort first and most heavily on the tenant and the lease, because that is where the financeability is principally determined, and should not let an impressive asset or a strong sponsor distract from a weak or short offtake. The asset cannot rescue a poor contract, but a strong contract can finance even a modest asset.

Implementation Considerations

Translating the framework into practice involves underwriting the covenant and lease first, structuring covenants and reserves, securing power and completion protections, and governing the loan over its life.

A foundational implementation point is that the lender should resist the pressure, common in a competitive, fast-growing market, to compress its diligence on the contract in order to win the deal. In a boom, sponsors have options and may push lenders to move quickly and to accept terms with less scrutiny, and a lender anxious to deploy may be tempted to truncate exactly the covenant and lease analysis the framework places first. This is precisely the wrong economy: the contract is what the lender is lending against, and abbreviating its assessment to win the deal is abbreviating the only analysis that determines whether the loan is sound. The disciplined lender holds to its contract-first diligence even under competitive pressure, accepting that it will lose some deals to less careful lenders, on the view that the deals it loses by insisting on a proper covenant assessment are often the ones it is better not to have made. Maintaining diligence discipline under deal pressure is the cultural counterpart to the analytical framework, and without it the framework is easily set aside at the moment it matters most.

A useful organising principle for the diligence is to structure it as a sequence of gates mirroring the pillars, so that the lender does not proceed to later, costlier diligence until the foundational questions are answered. The first gate is the covenant: is the tenant creditworthy enough to support the loan? The second is the lease: are its length and terms strong enough to secure the cash flow over the debt? The third is coverage: does the cash flow service the debt under combined stress? The fourth is the gating risks: are power and, for construction, completion secured and mitigated? A deal that fails an early gate is declined or restructured before the lender invests in the later, more detailed diligence, which both protects the lender and uses its underwriting resources efficiently. Sequencing the diligence as gates in priority order is the operational expression of the framework’s weighting, and it ensures that the most decisive questions are answered first, so that the lender’s effort is concentrated where financeability is determined rather than dispersed across a flat checklist.

Underwriting the Covenant and Lease First

The lender should begin its underwriting with the tenant covenant and the lease, because they determine financeability. This means assessing the tenant’s creditworthiness as rigorously as a corporate credit, the entity actually on the lease, not a parent or affiliate unless it guarantees, and scrutinising the lease for length, escalation, break rights, capex and repair obligations, and termination provisions. Only once satisfied that the contracted cash flow is strong and durable should the lender proceed to the asset, power and construction. Beginning with the contract ensures the lender does not invest diligence in an asset whose offtake cannot support the loan, and it anchors the credit decision where the analysis shows it belongs.

A specific structuring consideration for data centres is the treatment of the equipment and technology within the shell, which depreciates and must eventually be replaced. The building and power infrastructure are long-lived, but the IT equipment has a shorter life, and the lender should understand who bears the cost of refreshing it, the tenant under its capex obligations, or the owner, and how that cost affects the cash flow available to service the debt over the loan’s life. A lease that places the equipment-refresh obligation on the tenant protects the owner’s cash flow and the lender’s coverage; one that leaves it with the owner requires the lender to reserve for it. This is part of reading the lease terms in detail and aligning the loan structure to them, and it is a data-centre-specific refinement of the general principle that the lender must understand exactly which costs fall where, because every cost borne by the owner reduces the cash flow that repays the debt. Underwriting the capex and refresh obligations is therefore part of underwriting the coverage, not a separate technical matter.

Structuring Covenants, Reserves and Control

Concluding Comments

A data centre is, to a lender, only as good as the contract that pays for it. This paper has set out the underwriting model for a Gulf data centre, built around the offtake, the tenant covenant and the lease terms that determine financeability, the debt-service coverage that must be stress-tested, and the power and completion risks that must be mitigated as gating items.

The evidence and analysis support five conclusions. First, the tenant covenant and offtake dominate financeability, ahead of the asset. Second, covenant quality drives the leverage and cost of debt. Third, lease length, terms and covenants convert a building into a financeable, bond-like cash flow. Fourth, debt-service coverage must be tested under combined stress. Fifth, power security and completion risk are gating items that no coupon compensates for if unaddressed.

Before turning to the limitations, it is worth situating this paper within its cluster. The strategic companion paper establishes that the Gulf data-centre opportunity is large and that offtake and power are its dominant value drivers; this paper turns those drivers into a lender’s underwriting model, showing how the offtake and tenant covenant are assessed, how the lease and loan covenants protect the lender, and how coverage and the gating risks are tested. The further companion papers address the power and land bottlenecks the lender must diligence and the capital stack within which the debt sits. Read together, they take a lender from the strategic case through the underwriting of a specific asset to the structuring of its place in the capital stack. This paper’s contribution is the underwriting model, the disciplined, priority-ordered assessment of the contract that, to a lender, is what a data centre fundamentally is.

These conclusions are subject to limitations that also mark out the avenues for further work. The framework is stylised and illustrative, and a lender’s actual covenants, advance rates and coverage will differ; the value is in the structure, not the specific numbers. The analysis treats data-centre underwriting in general terms, whereas tenants, leases and power arrangements vary widely, and the technical assessment of a specific asset requires specialist diligence beyond this framework. The enforceability of the security and covenants depends on the legal regime, addressed in the companion paper on security and enforcement, and the financing structure on the capital stack, addressed in another companion paper. And the framework underwrites a single asset, whereas a lender building a data-centre book would also manage concentration and correlation across its loans. These are extensions rather than corrections: the central message, that data-centre underwriting is contract underwriting, dominated by the tenant covenant and the lease, disciplined by stress-tested coverage, and gated by power and completion, is robust. A lender that underwrites the covenant and lease first, structures protective covenants and reserves, secures power and completion, and monitors actively can lend against a Gulf data centre with confidence that its position rests on a secure, contracted cash flow, which is the foundation on which sound data-centre debt is built.

Disclosures