I've watched more AI infrastructure deals die in the gap between "promising site" and "closed capital" than from any other cause. The sites are real. The power is real. The opportunity is real. The deals don't close anyway.
The reason is consistent enough that I built an entire practice around it: developers and landowners speak a different language than institutional capital, and nobody owns the translation layer between them. What gets presented as a compelling opportunity reads, on the other side of the table, as a list of unpriced risks.
Capital doesn't walk away from risk. Capital prices risk — or it walks away from uncertainty. Those are different problems with different solutions. Understanding the distinction is the first thing that has to happen before a site can close.
The Five Most Common Deal-Killers
These aren't exotic edge cases. They show up in some combination on almost every deal I review.
01 — Power status is narrative, not evidence
Developers say "we have power access." Capital hears an ambiguous claim that could mean a letter of interest from a utility, a completed interconnection study, a signed service agreement, or something a landowner was told verbally eighteen months ago. These are not the same thing. The gap between a utility's letter of interest and an executed interconnection agreement can be two years and $40M in upgrade costs. When developers don't make that distinction explicit, capital assumes the worst version — because they've been burned by the optimistic version before.
What capital needs: the precise stage of utility engagement, the interconnection queue position, a realistic timeline to energization, and the cost to get there. Named. Documented. Not summarized.
02 — Permitting is "in process" with no milestone map
Permitting timelines in the data center space are genuinely variable — anywhere from six months to five years depending on jurisdiction, environmental constraints, and political climate. "In process" tells capital nothing about which end of that range they're looking at. They can't sequence their capital against a timeline they can't see.
What capital needs: a specific inventory of pending approvals, the critical path to a shovel-ready state, known objections and their status, and a probability-weighted timeline. If you don't have that, your permitting risk isn't priced — it's just hanging over the deal.
03 — Cooling and water constraints surface late
Cooling is the fastest-moving constraint in AI infrastructure right now. The liquid cooling requirements for dense GPU deployments are an order of magnitude beyond what air-cooled facilities needed five years ago. Water access, water rights, and municipal infrastructure capacity are limiting factors on sites that look clean until you go three levels deep into the diligence.
The problem isn't that constraints exist. The problem is that they surface in late-stage diligence, after capital has done preliminary underwriting and started to commit internally. At that point, a water availability problem isn't just a deal issue — it's a relationship issue. Capital doesn't forget which advisors brought them a site that wasted three months of due diligence.
04 — No capital sequencing architecture
Most site presentations describe a total capital requirement without explaining how that capital gets deployed, against what milestones, in what tranches, with what risk exposure at each stage. Institutional investors need to know what they're buying into. Is this early-stage development risk or operational cash flow? What does the capital stack look like? Who's in front of them? What triggers additional draws?
The absence of a capital sequencing plan forces each investor to build their own model of what they're buying — and those models will be conservative, because they're working from incomplete information. The developer loses control of their own narrative at the most important moment.
05 — The pitch is developer-centric
This is the hardest one to fix, because it requires developers to genuinely shift their perspective. A site presentation that leads with acreage, power availability, and location is written from the developer's vantage point. Institutional capital doesn't care about the asset in isolation — they care about the risk-adjusted return on their capital deployment, the path to liquidity, and the credibility of the development team's ability to execute.
A site with 200MW of available power and a clear interconnection path is still a bad investment if the development team can't demonstrate a credible path to an executed lease or purchase agreement with a creditworthy tenant. Capital is buying a probability-weighted projection, not a physical asset. The presentation has to be written in that language.
What Capital-Ready Actually Means
Capital-ready doesn't mean the site is perfect. It means the site's imperfections are known, named, and priced. An institutional investor can underwrite a site with real constraints — if those constraints are quantified. What they can't underwrite is a site where they don't know what they don't know.
The checklist version of capital-ready looks something like this:
- 01Power status documented to the specific stage of utility engagement, with timeline and cost clarity to energization
- 02Permitting inventory complete, with a critical path to shovel-ready state and known objections addressed or quantified
- 03Cooling and water diligence done upfront — not deferred to later-stage due diligence
- 04Capital stack designed before the conversation starts — tranche structure, milestone triggers, return architecture
- 05Tenant demand documented or credibly demonstrated — not assumed from market conditions
- 06Development team's track record presented in a form that speaks to the specific execution requirements of this site
Most sites that fail to close are missing two or three of these. The gaps aren't fatal — but unaddressed, they're deal-killers. Addressed correctly, they become the foundation of an investment thesis that capital can actually underwrite.
The Conversion Problem
The reason the gap persists is structural. Developers are experts in land, power, and construction. Capital allocators are experts in underwriting, return structure, and risk management. Neither side has deep fluency in the other's language. And there's no established professional category — no investment bank, no engineering firm, no law firm — that owns the translation layer between the two.
The result is a market where an enormous amount of real value is getting left on the table. Sites that could close capital don't, because nobody in the room can articulate the risk in the language capital needs to hear it. Capital that's ready to deploy sits on the sideline, because the deal flow that reaches them doesn't give them enough to underwrite.
This is the problem DCRE was built to solve. We bring a standardized assessment framework, institutional capital relationships, and the specific expertise to translate site reality into investment language. Not as advisors who hand you a report and walk away — as operators who stay in the deal until it closes.
The sites are there. The capital is there. The conversion problem is solvable. It just requires someone who understands both sides of the table well enough to sit at it.