Every data center conversation begins with energy, location, and capital, with power availability being the dominate subject matter. When real estate, capital markets, and power infrastructure collide, certainty around power is what dictates the entire deal structure.

The sheer magnitude of these projects has shifted just as dramatically.

Load additions that were historically measured in tens of megawatts are now campus requests seeking hundreds of megawatts, or even gigawatts, on accelerated schedules.

Industry analysis (Opens a new window) identifies two figures that frame the execution risk: Industry observers estimate speculative interconnection requests exceed real projects by a factor of five to 10.

• • Multi-year turbine backlogs and constrained equipment manufacturers capacity are limiting near-term buildouts regardless of how much capital is available.

Capital is not the binding constraint; sequencing is.

Structuring is the decision that shapes the rest

How you set things up at the onset drives what follows. We keep returning to that point because the downside risks for projects are largely a function of how they were structured in the first place.

Solving the power solution requires the analysis to determine the cost benefit analysis of owning the power generating assets, procuring an energy provider using distributed generation, or relying on the local utility for power and the different timelines involved for each of the solutions.

Resource adequacy concerns are pushing customers toward on-site distributed generation and near-site generation, including natural gas, combined heat and power, solar and renewables, battery energy storage, and thermal backup. Additionally, developers are evaluating the impact of demand flexibility on their ability to meet their overall energy needs.

When you own the generation, you own a tax question

For developers of data centers that own the power generating assets, even if they are bridging firm power from a utility, those assets could be generating tax credits and depreciation benefits. The time to model the impacts of those benefits is early in development – not after close.

Three paths are worth running side by side:

1. 1. Use the tax benefits and credits internally, against your own tax liability.
   2. Sell them to a third party, which is now possible under Section 6418.
   3. Consider the alternatives that may fit your position better.

Section 6418 opened a real market for many tax credits. Developers who can't fully absorb them against their own tax liability can now sell them outright to companies that can – generating immediate liquidity on one side and a lower tax bill on the other. Both parties take on documentation and compliance obligations, so the transaction needs to be structured carefully.

When a clean energy component becomes part of the power generation solution, those considerations layer-on further complexity.

Debt terms, capital stack reality, and the lender's view

The power generation solution for a data center asset directly shapes the financing conversation.

In higher-rate and shifting capital environments, developers may encounter tighter debt underwriting, including higher debt service coverage requirements, lower loan-to-value limits, and stricter covenant controls. Equity investors may also seek tighter sponsor controls, higher hurdle returns, and more extensive reporting requirements.

When modeling capital stacks for complex commercial and infrastructure assets:

• • Avoid relying solely on single-point return assumptions. Using both the DSCR method (from the lender's debt perspective) and the Band of Investment method helps establish a defensible capitalization rate range.
  • Test for negative leverage. When the cost of debt exceeds the going-in yield of the asset, equity returns erode quickly unless the revenue agreements are structured to offset the spread.
  • Plan for refinancing gaps. Changes in achievable loan terms create substantial refinancing shortfalls if exit or conversion timelines slip during construction; and,
  • Underwrite the credit, not just the real estate. Lenders do not want a substantial amount of risk (capital) tied to a lease with an operator that has a short lifespan.

Model the downside before you commit

Scenario planning on the front end mitigates the downside risks associated with structuring projects that are constructed to drive returns over 15 to 25 years.

The risks worth modeling are well documented:

• • Supply chain. Transformers, switchgear, turbines, and skilled electrical and construction labor are the critical lead-time items.
  • Commodity and tariff pressure. Volatile costs for materials, steel, and electrical equipment act as gating items for feasibility.
  • Policy movement. Shifting incentives, environmental rules, and permitting reform mean execution risk is now set as much by policy processes as by physical grid hardware.
  • Regulatory reassessment. States are changing the rules as we go, and even considering re-evaluating previously approved interconnection projects as demand projections shift.
  • Timing of firm power. Consider energy options if the timeline for utility delivered power is not met and potential consequences on the project.

Build flexibility into your contracts so the project doesn’t get hurt if government policies, tax incentives, or funding programs change during development. And, don’t bet the project on one government program. Keep multiple incentive options available so you can adjust/adapt if the rules change.

Evaluate the project on paper first

Using our industry experience across project finance, commercial real estate, construction and energy, we can help evaluate the project on paper before there are shovels in the ground and before the financing closes. Walking through the potential pitfalls, and what we are seeing in the industry today, gives clients an edge as they develop their projects.

Build the reporting package before you need it

You – and your lender – need to see where everything stands along the way. Evaluating the project from a risk perspective allows developers and owners to solve could-be problems before they impact a deal.

What that oversight should catch is specific. On data center builds, small-repeated errors compound: duplicate billing or billing for work not verified in the field, unauthorized change orders without approved scope documentation, skipped or postponed commissioning steps, equipment logged as delivered without factory acceptance testing, and missing medium- and high-voltage test reports that prevent safe energization.

The supporting infrastructure is where the exposure concentrates. Server halls are often close to standardized; substations, long-lead transformers, medium- and high-voltage distribution, cooling plants, water permitting, and utility interconnection capacity are not.

Catching issues in real time matters far more than running a flawless post-mortem. An end-of-project audit only catalogues what went wrong after the money is spent – active monitoring during construction lets you fix problems while you still have room to act.

Things to consider

• • Settle the power supply mix early in the project life-cycle and understand risks associated with the timeline for firm power.
  • Model all three credit paths as scenarios rather than relying on one.
  • Reconcile lender DSCR constraints with equity yield targets using dual cap-rate modeling methods.
  • Stress test against lead times and policy movement, not just baseline construction costs.
  • Bring in advisors while the structure can still change.

Let's talk early

If you are evaluating a data center project and the power and financing strategy is still taking shape, that is the useful moment for a conversation – not later.