The term “shovel-ready” no longer means much in data center real estate. What the market cares about is “power-ready” — and the delta between those two conditions is now the central variable in how acquisitions are priced, structured, and closed.
Interconnection risk has moved from a diligence footnote to the dominant underwriting question. Acquirers are modeling it explicitly. Sellers are marketing around it. Enterprises sitting on powered sites — facilities with firm utility commitments and clear interconnection pathways — are discovering they hold an asset the market is pricing at a premium it hasn’t commanded before.
This piece examines the mechanics of that shift: where the structural supply gap sits, how the regulatory environment is compounding it, why utilities have become deal co-architects, and what a disciplined grid risk framework looks like in an active transaction.
The 11 GW Problem: Announced Is Not the Same as Deliverable
The data center pipeline looks large on paper. Across 777 large data centers and AI factories announced since 2024, the industry is tracking approximately 190 GW of planned capacity. Of the 16 GW slated for 2026, only 5 GW is currently under construction. The remaining 11 GW is stuck in the announced stage — sites that exist on paper, with capital committed in principle, but without the utility coordination, interconnection agreements, or regulatory clearances required to break ground.
Sightline Climate’s data center outlook projects that 30 to 50 percent of the 2026 pipeline may not materialize. That is not a demand-side problem. Preleasing in US colocation markets sits at 81.5%, with Americas vacancy at 4.2% — effectively structural undersupply. The pipeline is stalling on the supply side, specifically on power access.
The interconnection queue is the mechanism through which that stall happens. Projects that reached operational status in 2025 spent an average of eight years waiting to connect to the grid, according to Novogradac’s analysis of queue data. In PJM, MISO, and CAISO — the three major interconnection regions covering most high-demand data center markets — wait times stretch four to five years even for well-capitalized developers with mature site control. For a transaction underwritten on a two-to-three year development timeline, that queue exposure is not a risk factor. It is a deal-killer dressed up in a pro forma.
This creates a bifurcation the market is now pricing explicitly: assets with secured, deliverable power on one side; assets with announced capacity and a queue position on the other. The gap between those two categories in deal multiples is widening.
The Regulatory Overhang: Moratoriums Are Not Hypothetical
Interconnection queue depth alone would be sufficient to reweight how acquirers think about grid risk. But the regulatory environment is adding a second layer of constraint that M&A teams have been slow to fully incorporate into their models.
As of early 2026, lawmakers in at least 11 states — including Georgia, Maryland, Michigan, New Hampshire, New York, Oklahoma, South Carolina, South Dakota, Vermont, Virginia, and Wisconsin — have introduced legislation that would temporarily ban or pause data center construction. The specific mechanisms vary: some bills pause permitting until utility studies are complete; others require that any new data center be co-located with dedicated on-site power generation sufficient to meet its own load. Good Jobs First tracked the spread of these proposals through February 2026, noting that the legislative posture is shifting from incentive-focused to regulatory-constraining across both red and blue states.
In Virginia — the largest data center market in the world — a proposed moratorium would halt new permitting until existing queue requests are fulfilled. New York is considering a three-year construction pause pending environmental and utility rate impact studies. Maryland’s proposal would condition any new development on co-located generation capacity matching the facility’s full load.
For M&A purposes, these proposals create material title risk on greenfield sites and expansion parcels. An asset acquired with modeled expansion capacity tied to a speculative interconnection application faces potential regulatory prohibition on that expansion pathway — a scenario requiring explicit stress-case modeling, not tail-risk treatment.
The practical effect is a legal bright line: existing, powered, operating assets are insulated from most of these proposals through grandfathering provisions. Development pipeline assets are not. That distinction is driving capital toward the former.
Utilities as Co-Architects: What This Means for Transaction Structure
Perhaps the most structurally significant shift in the market is the transformation of the utility relationship from passive infrastructure provider to active deal participant.
Duke Energy’s data center contract portfolio reached 4.5 GW — up from 3 GW following the Compass agreement — illustrating the scale of co-investment now underway. Microsoft’s Duke Energy arrangement is not a standard power purchase agreement. It represents a utility committing capital to upgrade transmission and distribution infrastructure in coordination with a specific operator’s development roadmap. The utility is a stakeholder in the project economics, not merely a service provider.
This co-architecture model changes what buyers need to underwrite. The relevant question is no longer whether a site has grid access at current capacity. It is whether the utility relationship supporting that site has the depth, contractual structure, and capital commitment required to support the buyer’s intended load expansion.
A site with 50 MW of operating load and a utility letter of intent for an additional 100 MW is not the same as a site with 50 MW of load and a fully executed interconnection agreement with engineering milestones and cost-sharing provisions. Both may appear equivalent in an offering memorandum. In structured diligence, they are categorically different assets.
According to S&P Global Market Intelligence, most acquisitions involve single-site operating assets with power already committed, or companies with established locations where utility relationships are in place. Sophisticated capital is concentrating on assets where the power question is already answered, not on development bets that require the buyer to navigate queue exposure post-close.
The Onsite Power Shift: Behind the Meter as a Structural Variable
The utility co-architecture trend runs parallel to a second structural shift: onsite and behind-the-meter power is moving from contingency infrastructure to primary architecture.
CoreSite’s 2026 outlook notes that data center operators are increasingly committing to behind-the-meter power arrangements — fuel cells, on-site gas turbines, battery storage systems, and small modular reactors in the planning horizon. The logic is straightforward: grid interconnection timelines are incompatible with AI infrastructure deployment timelines. A hyperscaler committing to 500 MW of new capacity for model training does not have four years to wait for a queue position.
The recent acquisition of a 132-acre property in Lufkin, Texas — including an existing power generation facility — valued the captive generation asset at $164 million against an $11.2 million site acquisition price. That ratio reflects how the market is currently pricing power certainty: the generation infrastructure is worth an order of magnitude more than the land, and their combination is the actual asset being acquired.
For transaction underwriting, this introduces a new line item in asset valuation: the captive generation capacity, its fuel security, its interconnection relationship with the broader grid, and its ability to operate islanded during grid disruption. These are not standard commercial real estate diligence items. They require engineering review, utility coordination expertise, and regulatory analysis of any behind-the-meter arrangements — all of which must be integrated into the deal timeline.
StackedAI’s Grid Risk Framework: Five Variables That Determine the Premium
Evaluating interconnection risk in a data center transaction requires a structured approach. The following five variables form the core of how StackedAI assesses grid risk across deal types — from single-asset acquisitions to portfolio carve-outs and enterprise campus dispositions.
1. Queue Position Depth and Vintage
Where does the site’s interconnection application sit in the regional queue, and when was it filed? A queue position filed in 2021 under pre-FERC Order 2023 rules carries different risk than a position filed in 2024 under the new financial readiness requirements. Vintage matters because rules, costs, and queue management practices have changed materially across ISOs in the past 36 months. Early-stage queue positions with limited deposit exposure are not equivalent to mature positions with completed feasibility studies.
2. Utility Relationship Formality
Is the utility relationship a letter of intent, a load growth study agreement, or a fully executed interconnection agreement with defined cost-sharing terms, engineering milestones, and a delivery date? Each of these represents a fundamentally different level of commitment — and a different risk profile for an acquirer assuming that capacity will be deliverable on the projected timeline.
3. Expansion Pathway Regulatory Exposure
Does the site’s modeled expansion capacity require new permitting in a state that has active moratorium legislation? Does expansion require additional interconnection applications, or does it operate under an existing agreement? What is the regulatory posture of the relevant state utility commission toward new large-load service requests? These questions are now required underwriting items, not optional considerations.
4. Onsite Generation Capacity and Fuel Security
Does the asset include captive generation? If so, what is the fuel source, what is the nameplate capacity relative to current and projected IT load, and what are the contractual terms for fuel supply? A site with 20 MW of diesel generation and 80 MW of grid-dependent load is not “power-redundant” in the current environment — it is 80% grid-exposed. Sites with firm behind-the-meter gas supply agreements, battery storage, or contracted renewable power purchase agreements with physical delivery from adjacent generation carry measurably lower operating risk.
5. ISO/RTO Market Structure
Interconnection risk is not uniform across regions. PJM, MISO, CAISO, ERCOT, and Southeast utility territories have materially different queue dynamics, cost allocation rules, and upgrade cost exposure. Enverus’s 2026 Interconnection Queue Outlook notes that project risk diverges sharply by region, and that developers and investors face growing pressure to assess this earlier in the process. ERCOT, with its independent grid and historically shorter queue timelines, commands a structural premium over PJM markets for development assets — a regional differential that should be reflected explicitly in deal underwriting, not absorbed into a generalized adjustment.
What This Means for Enterprise Holders of Powered Sites
The M&A implications described above are not limited to institutional investors evaluating colocation acquisitions. They apply directly to enterprises holding owned data center infrastructure with secured power.
The category of “enterprise data center with committed utility capacity” has become, structurally, an institutional-grade asset — regardless of whether the enterprise currently thinks of it in those terms. Campuses with 10 to 50 MW of contracted utility capacity, long-term interconnection agreements, and expansion headroom are exactly the kind of assets that private equity infrastructure funds and colocation operators are actively sourcing. The combination of scarcity (4.2% Americas vacancy, 11 GW of pipeline stuck in the announced stage) and regulatory compression (11 states with active moratorium proposals) means that existing, powered, operating sites are in a supply category that cannot be replicated on a short timeline.
For an enterprise evaluating its real estate position, this creates three actionable considerations. First, the utility documentation supporting any owned data center facility — interconnection agreements, service agreements, capacity commitments — should be audited for completeness and transferability. Institutional buyers require clean utility records as a condition of any transaction. Second, unused but contractually preserved expansion capacity has real option value that should be reflected in any disposition analysis. Third, timing is material: the regulatory window for executing a disposition at premium pricing is tied to how long the supply-demand imbalance persists. If even a fraction of the 11 GW announced-stage pipeline converts to operational capacity over the next 18 to 24 months, the scarcity premium compresses.
The window is not permanent. But it is open.
Advisory Conclusion
Grid scarcity has restructured the data center M&A market’s risk hierarchy. Power access — its certainty, its formality, its regulatory durability — is now the primary variable in asset valuation, ahead of lease structure, tenant credit, and even geographic positioning in most active deal processes.
The 113 transactions completed globally in 2025, representing more than $69 billion in total deal value, were disproportionately concentrated in assets where the power question was already resolved — institutional capital pricing certainty at a premium and demanding a discount for queue exposure.
For buyers, the framework is clear. Interconnection risk must be modeled explicitly, not absorbed into a generalized development risk adjustment. The five variables above — queue vintage, utility formality, regulatory pathway, onsite generation, and ISO market structure — form the basis of a defensible risk-adjusted underwriting model for any asset where power access is not fully contracted and operating.
For enterprise holders of powered sites, the market is sending a signal that is worth taking seriously. The assets you built for operational continuity may be worth considerably more to an infrastructure acquirer than they are to your balance sheet. The question is whether you have the market intelligence and transaction readiness to act on that asymmetry before the window closes.
StackedAI advises enterprise and institutional clients on data center M&A strategy, grid risk evaluation, and powered site disposition. Contact our advisory team to discuss your portfolio position.