Value
Financeability
Power source
Electron Economics · AI Infrastructure Intelligence

Intelligence Overview

Decision-grade signals on powered land, energisation certainty, and capital cycle risk across 61 transactions 2015–2026. All primary research — press releases, SEC filings, utility IRP filings, FERC dockets.

Market Pulse
Largest Live Risk
Conditional Capacity
Stargate + Aligned — pipeline MW without filed interconnection
$140B
2 Conditional deals · 10,000 MW announced
View Conditional deals
Most Bankable Scarcity
Secured Renewable Power
atNorth 1 GW Nordic · Sabey Grant County PUD · Crane Nuclear PPA
Power Risk 1
3 deals — grid bypass or secured renewable
View Scarcity Map
Active Seller Watch
Hold Duration Pressure
CyrusOne (KKR/GIP, 4yr) · Evoque (Brookfield, 7yr) · Vantage US (Macquarie, 2yr)
3 assets
Seller score 3+ · PE exit window approaching
View Potential Sellers
Latest Signals
May 2026 · Manually curated · Updated with each dataset revision
atNorth closes at $4.2B — ~9x in 4 years, 100% driven by Nordic renewable power repricing New deal
Equinix + CPP Investments · Feb 2026 · Partners Group exit · Power Risk 1 · Bankable
Aligned $40B closes H1 2026 — operating campuses Bankable, 3,600 MW pipeline without filed interconnection Watch
AIP / BlackRock GIP / MGX · Oct 2025 · Power Risk 4 · Conditional · $28.6M/delivered MW
KKR + Singtel acquire STT GDC for $10.9B — 1.7 GW APAC operating platform, established SP Group connections Updated
Feb 2026 · Power Risk 1 · Bankable · $6.4M/operating MW · APAC multi-market
Core Scientific shareholders reject CoreWeave $9B bid — $6.9M/MW below clearing price for ERCOT HPC capacity Signal
Oct 2025 · Power Risk 1 · Stalled · Sets floor: established energised ERCOT capacity >$7M/MW
SoftBank acquires DigitalBridge for $4B — $108B AUM, management economics on Switch, Scala, DataBank, Vantage Updated
Dec 2025 · Platform acquisition · Expected close H2 2026 · Marc Ganzi continues as CEO
Featured Deal Briefs
⚡ Powered Land Scarcity Map — click to explore
Open full map →
Bubble area ∝ transaction value · Color = power risk · Click to open full interactive map with deal filtering
61 transactions · 2015–2026 · Primary research · No third-party data subscriptions
All signals and briefs are primary research assessments based on publicly identifiable filings as of the dataset vintage date. Financeability tiers reflect grid access, offtake, and construction finance conditions at announcement. Not investment advice. Verify independently before use. · electroneconomics.substack.com
Electron Economics · Scarce MW Intelligence

Where can you still acquire bankable power?

51 major AI infrastructure transactions 2015–2026, mapped through a single lens: financeability. Not which deals happened — which megawatts are scarce, bankable, and likely to transact next.

⚡ Powered Land Scarcity by ISO Territory — click to explore full map
Open full map →
Bubble size ∝ transaction value · Colour = power scarcity level · Green = low scarcity / available · Red = critical / constrained
Most Bankable Scarcity Positions
Established grid · Lowest replication risk
Conditional MW Watch
Announced but not yet bankable — grid delivery unresolved
Reality Check — Lowest Execution Probability
Announced capacity with unresolved delivery risk
Land Grab — Top MW Acquirers
Total MW controlled by buyer 2015–2026
61 transactions · 2015–2026 · Primary research · electroneconomics.substack.com
Markets · Land Grab

Who's Winning the Powered Land Grab?

Total megawatts controlled by each acquirer across all disclosed transactions 2015–2026. Energised MW = operating today. Total MW includes contracted and pipeline capacity.

Primary research · All MW figures from disclosed transaction terms · Pipeline MW includes announced but not yet energised capacity · electroneconomics.substack.com
Intelligence · Changelog

What's changed

Every deal addition, status update, and platform change — full primary research audit trail. Electron Economics maintains a living dataset; this log documents every material revision since launch.

Electron Economics · AI Infrastructure · Primary Research

AI Infrastructure Transaction Intelligence 2015–2026

Who is acquiring control of future powered capacity, at what implied $/MW, and with what execution risk? 61 transactions (2015–2026) classified by deal logic, power risk score, and financeability tier — all primary research from SEC filings, FERC dockets, utility IRP filings, and press releases.

V4 · 51 DEALS 2015–2026 · 51 static briefs
⬤ SEC / regulatory filing
◎ Press release
○ Reported / inferred

The governing question: who is buying control of future deliverable megawatts, at what implied $/MW, and with what execution risk? 44 of 61 transactions ($206B) were Bankable at announcement — firm grid, signed offtake, construction path. The two Conditional deals (Stargate + Aligned) represent $140B and 10,000 MW. They are also the two deals with the most uncertain power delivery paths. Every data center deal is a power delivery deal.

Total Deals
in dataset
Total Value
USD disclosed
Bankable
firm grid + offtake
Conditional
grid pending
Energised MW
commissioned IT load
Pipeline MW
announced, not energised
Saved views:
Compare
Select 2–4 deals
Deal Comparison
Status:
Financeability:
Power:
Logic:
Date Asset / Deal Type Buyer Value ($B) MW $/MW ($M) Power Source Grid Access ISO / Utility Deal Logic Power Risk Financeability Tenant Execution Status Conf.
Transaction Value
USD at announcement
IT Load (MW)
operating unless noted
Implied $/MW
USD millions
Date
announced / closed
Tenant Type
primary occupant
⚡ Power & Grid Intelligence
Power Source
Grid Access Status
ISO / Utility Territory
Financeability
All power intelligence fields sourced from primary records: press releases, SEC/FERC filings, utility IRP filings, and ISO queue public data. Financeability tier reflects primary research assessment of whether grid access, offtake, and construction finance conditions were met at announcement — not current status. MW = IT load (operating) for M&A; announced pipeline for Campus/Power deals.
Electron Economics · Primary Research

Power Risk Intelligence

Every data center deal is a power delivery deal. This view reframes the transaction record through one lens: what was the power delivery status at announcement, and who wore the grid access risk?

The governing thesis. Capital is not the binding constraint in AI infrastructure. De-risked, energised, bankable megawatts are. A campus without a firm interconnection agreement and a credible energisation schedule cannot be financed — not because lenders are squeamish, but because without a delivery date there is no lease commencement date, and without that there are no contracted cash flows to lend against.

The financeability tier below reflects one primary research question applied to each transaction: at announcement, did the deal have (1) a firm interconnection agreement or equivalent grid bypass, (2) a signed offtake or anchor lease, and (3) a construction finance path? All three → Bankable. Missing one → Probable. Missing two → Conditional. Missing all three or development-stage announced capacity → Pre-finance.

The nuclear PPA pattern. Amazon's 960 MW Crane Clean Energy Center PPA and Microsoft's 10.5 GW Brookfield PPA are not sustainability plays. They are financing manoeuvres: nuclear and long-duration renewables already have interconnection and a delivery date. The PPA premium is less about carbon than about schedule certainty — which is what converts announced MW into bankable MW.

Latest Dataset Updates
May 2026 · Maintained with each primary research cycle
STT GDC / KKR / Singtel — $10.9B, 1.7 GW APAC — added to dataset New deal
Feb 2026 · Power Risk 1 · Bankable · Established SP Group / TEPCO / Kansai connections · APAC multi-market
Aligned Data Centers — Financeability revised to Conditional; pipeline MW reclassified Revised
Oct 2025 · $40B · Power Risk 4 · Operating campuses Bankable; 3,600 MW pipeline lacks filed interconnection
atNorth (Equinix / CPP) — $4.2B, ~9x in 4 years — deal brief added Brief added
Feb 2026 · 1 GW secured Nordic power · Landsvirkjun geothermal / Vattenfall hydro · Power Risk 1
Core Scientific shareholder vote — CoreWeave $9B offer rejected; implied floor for ERCOT HPC capacity established above $7M/MW Signal
Oct 2025 · Execution: Stalled · Sets price floor for energised US HPC data center capacity · Power Risk 1
PJM Dominion scarcity notes updated — queue lead times revised to 30–42 months Scarcity
Q1 2026 · Virginia SCC rate case filings · Dominion interconnection queue processing rate declining under record load applications
Deal Value by Power Source ($B)
MW Capacity by Power Source
Deals by Financeability — click to filter  
Financeability tiers are primary research assessments based on public filings and press releases at announcement date, not current status. Grid access status inferred from interconnection queue filings, utility rate case disclosures, and press releases.
Electron Economics · Chronological

Deal Timeline

All transactions ordered by date. Color: green=closed, purple=pending/announced, red=canceled. Click any entry to open detail.

● Closed
● Pending / Announced
● Canceled
Markets · Analytics

Visual Intelligence

Five analytical exhibits drawn from 61 transactions 2015–2026. Each answers a question practitioners actually ask — not what happened, but what it means for financeability, scarcity, and exit timing.

Exhibit 1 · Financeability Matrix
Power risk vs. $/delivered MW — Bankable deals cluster; Conditional deals are outliers
Bubble = deal value. Colour = financeability tier. The $11–12.5M/MW Bankable cluster is the market's clearing price for energised infrastructure.
Source: Electron Economics primary research · Delivered MW = Energised + Contracted · Excludes pipeline MW without filed interconnection
Exhibit 2 · Capital Deployment
$324B in 11 years — the AI era 2024–2026 represents 73% of total disclosed value
Annual disclosed deal value. The 2021–2022 take-private supercycle ($52B) is dwarfed by the AI era acceleration.
Source: Electron Economics primary research · SEC filings, press releases 2015–2026
Exhibit 3 · Exit Pressure Matrix
Hold duration vs. exit probability — three assets are ripe for transaction
X axis = hold years since acquisition. Y axis = seller score (1–5). Top-right quadrant = high conviction exit candidates. Bubble = entry value.
Source: Electron Economics primary research · Seller score = proprietary assessment of exit likelihood within 36 months
Exhibit 4 · Deliverable MW Analysis
22,850 MW in pipeline — most is not yet bankable; only energised MW can be financed today
MW breakdown by delivery status and financeability tier. The gap between announced and deliverable is where underwriting risk lives.
Source: Electron Economics primary research · Pipeline MW = announced without filed interconnection application
Exhibit 5 · Transaction Universe
51 deals mapped by value vs. MW — power source drives the premium; Stargate and Aligned are structural outliers
X axis = total MW. Y axis = transaction value ($B). Bubble size = implied $/MW. Colour = power source. Dashed line = $12M/MW Bankable benchmark. Deals above the line paid a premium to replacement cost.
Source: Electron Economics primary research · 61 transactions 2015–2026 · Bubble size ∝ implied $/MW blended · Dashed line = $12M/MW Bankable market benchmark
Section 2 · Deal Structure & Composition
Exhibit 6 · Deal Logic Mix
Take-privates and platform recaps account for 79% of disclosed value — PPAs are structurally distinct instruments
Disclosed value ($B) and deal count by logic type. Hyperscaler PPAs are off-balance-sheet commitments, not equity transactions.
Source: Electron Economics primary research · 61 transactions 2015–2026
Exhibit 7 · Power Source Premium Heatmap
Nuclear assets command 3× the $/MW of grid-dependent deals — scarcity is priced into the molecule
Avg $/delivered MW by power source × financeability tier. Cell colour = relative premium. Numbers = avg $M/MW · deal count.
Source: Electron Economics primary research · Delivered = Energised + Contracted MW
Section 3 · Market Geography & Scarcity
Exhibit 8 · ISO Territory Scorecard
PJM Dominion and CAISO are critical-scarcity; ERCOT and Nordic markets retain relative availability
Per-territory: deals tracked, total MW, avg power risk, avg $/delivered MW, queue status. Click row to filter tracker.
Source: Electron Economics primary research · Queue status = primary research assessment · FERC queue + utility IRP filings
Section 4 · Deal Sizing & Distribution
Exhibit 9 · Transaction Size Distribution
AI era created a bimodal market — bolt-ons below $3B and mega-deals above $10B dominate; the mid-market has hollowed out
Deal count by transaction value bucket. Colour = era. The $10B+ bucket is an AI-era phenomenon.
Source: Electron Economics primary research · Excludes undisclosed values · n=52
Exhibit 10 · Implied Return Proxy
Short hold + low entry $/MW = maximum upside; long hold + high $/MW signals refinancing pressure by 2028
X = hold years since acquisition. Y = implied blended $/MW at entry. Bubble = deal value. Bottom-left = highest return potential.
Source: Electron Economics primary research · Hold = years since close date
Section 5 · Exit & Risk Intelligence
Exhibit 11 · Risk-Exit Matrix
Three assets in the danger zone: high power risk AND high exit probability — potential forced sellers at a discount
X = power risk (1–5). Y = seller score (1–5). Top-right = forced/distressed. Bottom-left = stable sovereign. Bubble = entry value.
Source: Electron Economics primary research · Seller score = proprietary exit likelihood assessment
Exhibit 12 · Tenant Mix by Era
AI Cloud (CoreWeave, xAI, OpenAI) is a 2024–2026 phenomenon — didn't exist as a buyer category before this cycle
Deal count by tenant category and era. Enterprise dominated colo-era. Hyperscalers absorbed the take-private wave. AI Cloud emerged 2024.
Source: Electron Economics primary research · Tenant classification = primary research assessment at announcement
Exhibit 13 · MW by Power Source
Renewables dominate by MW volume; Nuclear MW commands 4× the $/MW premium
Total MW tracked by power source (energised vs total). Inner ring = energised; outer = all MW including pipeline.
Source: Electron Economics primary research
Exhibit 14 · Capital by Buyer Category
PE and infrastructure funds deployed 68% of disclosed equity — sovereign wealth is the silent senior anchor
Disclosed value by buyer type. Hyperscaler PPAs excluded. Classification = primary research by entity type at announcement.
Source: Electron Economics primary research · Buyer classification based on disclosed entity type
Exhibit 15 · Reality Check Score Distribution
Bankable deals cluster 80–95%; Conditional deals (Stargate 42%, Aligned 51%) are structural outliers, not a continuum
Execution probability for all 61 deals. 6 factors: Interconnection (25pts) · Offtake (25pts) · Financing (20pts) · Land (15pts) · Generation (10pts) · Execution (5pts). Hover = deal name. Click = brief.
Source: Electron Economics primary research · Reality Check = proprietary execution probability framework · Not investment advice
Electron Economics · Geographic Intelligence

Powered Land Scarcity Map

Transaction concentration, MW volume, and power risk by ISO territory and utility zone. Bubble size = transaction value. Color = power risk score. Click any ISO card to filter the deal list.

Deals
Value
MW (energised)
Avg Power Risk
Scarcity Signal
Risk 1–2 (Low)
Risk 3 (Moderate)
Risk 4–5 (High)
Bubble area ∝ transaction value
Deals by ISO Territory — Filtered to selected ISO if active
ISO territories approximated from deal-level utility disclosures. Bubble positions are schematic, not parcel-accurate. Scarcity signals based on primary research from FERC queue data, utility IRP filings, and press releases.
Electron Economics · Forward Intelligence

Potential Sellers Watchlist

Operators scored on likelihood of coming to market within 36 months, based on sponsor hold duration, leverage profile, occupancy trajectory, power risk, and strategic review signals. All primary research assessment — no third-party data.

Scoring methodology. Seller score 1–5 is a primary research judgment combining four factors: (1) hold duration — PE sponsors typically target 3–7 year holds; assets beyond 5 years are under natural exit pressure. (2) financeability — Conditional assets face refinancing friction that accelerates sale timing. (3) execution slippage — delayed energisation reduces LP patience. (4) market comparables — recent comp transactions set a mark that motivates seller crystallisation. Score 3+ warrants active monitoring.

Disclaimer. This is analytical inference from public signals — hold duration, deal vintage, transaction structure, and power risk — not insider information. Treat as a hypothesis, not a prediction.

Seller Score Distribution — by Hold Duration
Seller scores are inferred from public deal data: sponsor hold duration, financeability tier, execution slippage, and market context. Not investment advice.
Electron Economics · AI-Powered Analysis

Deal Brief Institutional research · Primary sources

Select any transaction to read an institutional-grade deal brief: strategic significance, power delivery analysis, financing implications, comparable transactions, and structural observations. All 51 briefs are pre-authored primary research — no API required.

Select Transaction
Deal Brief · Electron Economics
Primary research analysis by Electron Economics. All figures sourced from press releases, SEC filings, utility IRP filings, and FERC dockets. Not investment advice. Verify independently before use.
All 51 deal briefs are pre-authored primary research. Sourced from press releases, SEC/FERC filings, utility IRP filings, and ISO queue data. No third-party data subscriptions.