Most data center projects don’t fail because someone missed a box on a checklist. They fail because the checklist was the whole strategy.
Data center due diligence is the pre-investment process of evaluating whether a site can actually be built, covering power availability, regulatory and policy risk, community acceptance, physical infrastructure, and capital structure. Done well, it doesn’t just assess each area in isolation. It stress-tests how they interact, because that’s where projects break.
Right now, the appetite to build is enormous. AI and high-density compute are pushing capacity requirements past anything the last decade prepared us for, and the capital is lining up to meet it. On paper, the math looks straightforward: secure the land, line up the power, clear the permits, and start pouring concrete. In practice however, the projects that look cleanest on paper are often the ones that stall hardest once they hit the real world.
The reason is almost never a single fatal flaw. It’s the gaps between the pieces. And conventional due diligence, the kind that hands each domain to a different specialty firm and staples the reports together, is built to miss exactly those gaps.
What does data center due diligence need to cover?
Walk through what a developer actually has to understand before committing capital, and the scope becomes clear fast.
There’s power, which is now the defining constraint on almost every large project. Access to power, not land or fiber, sets the pace. Interconnection queues are backed up: the typical project reaching commercial operation now spends more than four years in the queue, and roughly 2,290 GW of capacity sits active across U.S. queues, nearly twice the entire installed U.S. fleet, according to Lawrence Berkeley National Laboratory. Study phases can run for years, and demand charges and tariff structures for large loads can reshape the economics of a site long after the ink is dry.
There’s regulatory and policy risk, which shifts by state, by utility, and by proceeding. The rules governing large-load interconnection are being rewritten in real time. In 2025 alone, more than 200 data center bills were introduced across 40-plus states, and the pace accelerated into 2026, with a growing number of states now weighing outright construction moratoriums. What’s permissible in one market may be actively contested in the next.
There’s community engagement, which developers routinely underweight until a zoning hearing goes sideways. Local sentiment for water use, noise, tax abatements, and land use can accelerate a project or bury it. There’s physical infrastructure, from substation capacity to water and fiber, each with its own lead time and its own set of owners.
And there’s capital, which too many teams treat as the solvent for all of the above. The assumption is that with enough money and the right deal structure, the other constraints bend. They don’t. Power complexity is structural. A fragmented system of utilities, regulators, and regional markets doesn’t move faster because a project is well funded.
Here’s what makes this genuinely hard: none of these domains sits still while you evaluate the others. A demand-charge tariff that looks manageable in isolation can turn a site uneconomic once you layer in an interconnection timeline that pushes energization out two years. A community that would have welcomed the project in year one may sour by the time permits are ready. The power plan and the permitting plan and the capital plan are not independent workstreams. They’re one system, and they fail as a system.
Why can’t the standard specialty-firm approach catch these risks?
The common way to run diligence is to parcel it out. A real estate firm handles the site. An energy consultant handles the power study. Outside counsel handles the regulatory read. Each comes back with a clean report inside its own lane.
The trouble is that risk doesn’t respect those lanes. The power consultant signs off on interconnection feasibility without knowing that the tariff assumptions depend on a load ramp the community engagement team already flagged as unrealistic. The land attorney clears zoning without seeing that the substation upgrade required to make the site viable sits on a five-year utility timeline. Every report is correct in isolation. The project still fails, because no one was accountable for how the pieces stress each other.
A no-BS approach to due diligence starts from the opposite premise. The point of diligence is to find where the project breaks, and projects break at the seams. That means pressure-testing the interdependencies directly: what happens to the capital plan if energization slips a year, whether the tariff still works under a slower ramp, how a permitting delay cascades into the interconnection schedule. It means telling a developer that a site pencils out only if three separate things go right at once, and being honest about the odds of that. It means being willing to say a deal doesn’t work before the capital is committed rather than after.
Who needs to be involved in data center due diligence?
This is why the question of who runs your diligence matters as much as what they look at.
You need real depth in each domain. Someone who genuinely understands interconnection queues and cost allocation, someone fluent in the specific regulatory proceedings that govern a given market, someone who has sat through the community hearings and knows how they actually go. Surface-level generalists miss the details that sink projects, and details are where these projects live.
But depth alone isn’t enough, because the specialists have to be in conversation with each other. When the power expert, the policy expert, the real estate lead, and the infrastructure team work under one roof, the interdependencies surface early, while they can still be designed around. When they work at four different firms, the gaps only surface after they’ve become problems, usually when it’s expensive to fix them.
That’s the model ASG is built on. Real estate, policy, power, and infrastructure expertise in one place, working the same project together, so the diligence reflects how the project will actually behave instead of how it looks in four separate PDFs.
How do you get started, or move through a bottleneck?
If you’re early, start with the constraint most likely to kill the project, which is almost always power, and work outward from there. Build the diligence around the interdependencies from day one rather than treating power and permitting and capital as sequential steps to clear.
If you’re already in motion and stuck, the bottleneck is rarely where it appears. A permitting delay may really be a community problem. A financing gap may really be an unresolved interconnection question that makes the timeline uninvestable. Diagnosing the real bottleneck usually means looking at the project as a whole rather than escalating pressure on the one workstream that happens to be visibly stalled.
Either way, the projects that get built are the ones whose teams took the interdependencies seriously before breaking ground, partnered with utilities and regulators instead of trying to route around them, and were honest with themselves about what could actually be delivered.
If you’re weighing a site, or trying to work through a bottleneck on a project already underway, ASG can help you see the whole picture before it costs you.
Reach out to start the conversation.
Frequently asked questions
What is data center due diligence?
Data center due diligence is the pre-investment process of confirming that a site can actually be built, not just that it looks viable on paper. It evaluates power availability and interconnection timelines, regulatory and policy risk, community acceptance, physical infrastructure, and capital structure, and it tests how those areas interact rather than assessing each one in isolation.
Why do data center projects fail diligence?
Most projects fail at the seams between domains rather than on a single flaw. A power study, a zoning review, and a financing model can each look sound on their own while the combination doesn’t hold, for example when an interconnection delay pushes energization past the point where the tariff or the capital plan still works. Fragmented diligence, split across separate specialty firms, is built to miss those cross-domain gaps.
How long does data center due diligence take?
It varies by market and site complexity, and the binding factor is usually power. Interconnection studies alone can run for years, and the typical project reaching commercial operation now spends more than four years in the queue, so serious diligence starts with time-to-power and works outward. The evaluation itself can move quickly, but it has to account for timelines measured in years, not quarters.
What makes a data center site fail?
The most common killers are power that can’t be delivered on a workable timeline, a regulatory or policy shift such as a moratorium or a large-load tariff change, community opposition that stalls permitting, or infrastructure gaps like insufficient substation capacity. Any one can sink a site, but the harder failures come from these factors compounding, which is why they have to be evaluated together.
Who should run data center due diligence?
Diligence needs genuine depth across power, policy, real estate, and infrastructure, and those specialists have to work in conversation with each other rather than in separate reports. When the experts share the same view of a project, the interdependencies surface early enough to design around. When they’re split across four firms, the gaps tend to surface only after they’ve become expensive problems.
