AI Capex Absorption Score
Quantifies whether a region or grid can absorb a hyperscaler buildout in dollars, megawatts, megaliters, and engineers.
Problem solved
Hyperscalers announce gigawatt-scale data center commitments faster than any single region's grid, water, talent pipeline, and permitting capacity can keep up. ACAS scores absorption probability for a named site against a four-axis stack (power, water, talent, permitting) and produces a delivery-risk classification with a gap-to-COD timeline and an explicit mitigation table.
Inputs
- Power available and queue position (utility filings, ERCOT, PJM, CAISO, MISO interconnection trackers)
- Water rights and aquifer stress (USGS, EPA WaterSense, regional water authority data, state engineer filings)
- Engineering talent depth (BLS QCEW occupational, IEEE membership density, regional university pipelines)
- Permitting throughput (state and county filings, prior data center COD-to-permit timelines)
- Transformer slot reservation and lead time (utility procurement, GE Vernova, Hitachi Energy, Siemens Energy disclosures)
- Transmission upgrade timeline (RTO planning queue, FERC Order 2023 cluster studies)
- Tax incentive package (state and county MOU, abatement, sales tax exemption)
- Community opposition signal (FERC and PUC docket comments, local press coverage, council vote tallies)
Outputs
- Delivery-risk class (low, medium, high) for the named site
- Individual axis scores (power, water, talent, permitting) on a 0 to 100 calibrated band
- Gap-to-COD timeline (months to commercial operation date) under base, stretch, and stress
- Mitigation ranking by axis with capex and timeline cost
- Scenario sensitivity to ERCOT-style large flexible load reform, FERC Order 2023 cluster timing, and water curtailment
- Behind-the-fence generation feasibility flag (gas peakers, on-site solar, SMR optionality)
- Replication package with named source links and rebuild scripts
Method
- Step 1. Lock the named site and its rated load. Resolve net vs. gross MW, ramp profile, and PUE assumption.
- Step 2. Pull power axis: queue position, transformer slot reservation, transmission upgrade timeline, large-load tariff exposure.
- Step 3. Pull water axis: water rights, aquifer status, cooling architecture (closed-loop, evaporative, dry), reclaimed-water availability.
- Step 4. Pull talent axis: engineering depth (electrical, mechanical, controls, fiber, SCADA), regional commute geography, hyperscaler talent saturation.
- Step 5. Pull permitting axis: state and county throughput, prior data center timelines, environmental review tier, community-opposition signal.
- Step 6. Score on a 0 to 100 calibrated band per axis, then combine with explicit weights (default 35 power, 25 water, 20 talent, 20 permitting).
- Step 7. Run interconnection queue stress and mitigation table; output replication package with citations.
Assumptions
- ERCOT, PJM, CAISO, MISO interconnection queue data is current as of the run date.
- Water rights are taken at face value from state engineer filings; pending litigation is flagged but not adjudicated.
- Talent pipeline is regional QCEW; cross-state recruitment is not modeled as a base-case mitigation.
- Default axis weighting is 35 power, 25 water, 20 talent, 20 permitting; users can re-run at custom weights.
- Behind-the-fence generation is treated as a mitigation, not a base-case input.
Limitations
- The framework does not model behind-the-fence generation in the base case; it appears only in mitigation.
- Community opposition is a signal, not a probability; the framework reports the signal but cannot resolve the political outcome.
- The four-axis weighting is debatable. CAISO and ERCOT users typically reweight power higher; arid-region users typically reweight water higher.
- Transformer lead-time data is procurement-confidential in many cases; the framework uses public OEM disclosures as a floor.
- FERC Order 2023 cluster-study timing is RTO-specific and continues to evolve.
Example application
Applied to a hyperscaler siting decision in ERCOT 2026: a 1.2 GW campus near Abilene scores high power axis (queue position favorable, transformer slot reserved), medium water axis (Edwards-Trinity aquifer stress flagged), low talent axis (regional engineering depth thin without commute reach to Dallas-Fort Worth), and medium permitting axis (county throughput recent but environmental review tier escalates with cooling architecture). The composite delivery-risk class is medium, with talent and water as the binding axes and a 14 month gap-to-COD under base case. See Texas ERCOT AI siting in 2026: power, water, and the absorption ceiling.
Where the method has been applied.
Texas, ERCOT, and the AI Siting Reset
Senate Bill 6, the ERCOT large-load study, and a 30 to 40 GW interconnection queue are forcing hyperscalers to rethink West Texas, behind-the-meter gas, and the...
Read brief → 2026-04-25Quebec hydropower and the new gating of AI compute
Quebec spent two decades selling itself as the cheapest, greenest place on the continent to plug in a data center. In 2026 Hydro-Quebec is throttling new connec...
Read brief → 2026-04-26Johor's Data Center Boom: Singapore's Spillover and Malaysia's Grid Bet
Singapore's moratorium pushed roughly 1.6 GW of latent demand across the causeway, and Johor is now Southeast Asia's most concentrated hyperscaler buildout. The...
Read brief → 2026-04-25AI capex met the grid: when the megawatt curve breaks
Hyperscaler capital spending crossed 500 billion dollars across 2025 and 2026 while the average US interconnection wait sits above 4 years. The constraint is no...
Read brief → 2026-04-26The 2026 Grid Capex Supercycle: Wires, Transformers, and the Cost of Connection
US transmission and distribution capex has tripled in a decade, EU TYNDP commits EUR 600 billion through 2034, and the binding constraint has moved from generat...
Read brief → 2026-04-26PJM capacity market in 2026: what the next auction is telling us
After the 2025-26 delivery year auction shocked the market with a more than ninefold price jump, the upcoming PJM capacity auctions will determine whether the l...
Read brief → 2026-04-26Hyperscaler GPU Procurement 2026: H200 vs B200 vs GB200 in Honest Deployment Math
Blackwell is no longer a roadmap promise, it is a procurement reality, and the only honest comparison runs on workload-weighted utilization rather than peak FLO...
Read brief → 2026-04-26European AI sovereignty in 2026: Mistral, Aleph Alpha, Stargate UAE, and the regulatory framework war
Europe spent 2024 and 2025 trying to build a sovereign frontier through Mistral and Aleph Alpha while the United States staged a 500 billion dollar Stargate bui...
Read brief → 2026-04-26The grid wakes up: AI demand, FERC 1920, and the PJM, MISO interconnection cliff
After two decades of flat US electricity demand, AI driven data center load is forcing a transmission and capacity build that the queue, the auctions, and the c...
Read brief →Related methods.
FEOC Stack
Foreign-entity-of-concern decomposition for IRA Section 30D and 45X bills of materials.
Pass-Through DecompTariff Pass-Through Decomposition
Five-step decomposition of who pays a tariff: importer margin, exporter price, or domestic consumer.
Compute CurveCompute Cost Curve
Six-step GPU TCO decomposition across acquisition, power, cooling, network, depreciation, and idle.
Multiplier BenchFiscal Multiplier Bench
Country-level fiscal multiplier estimates by spending category, business cycle phase, and monetary regime.
AI Siting ScoreEnergy-AI Siting Score
Composite jurisdiction ranking for AI compute siting across firm power, water, latency, talent, and policy stability.
IPMIIndustrial Policy Maturity Index
Five-dimensional score of whether a country's industrial policy is investible, across design, disbursement, conditionality, monitoring, and political durability.
Substitution MapTariff-Substitution Elasticity Map
Estimating which third-country supplier captures share when a tariff hits the primary origin, by HS6, with confidence bands.
Restructuring StackSovereign Restructuring Stack
Ordering of sovereign debt instruments by haircut tolerance, legal seniority, and political optics.
CMCICritical Minerals Concentration Index
HHI-based concentration scoring on extraction, processing, and refining for any critical mineral chain.
SDPSSovereign Default Probability Stack
Six-factor early-warning framework for emerging-market sovereign credit, calibrated to the post-Common Framework dataset.
EOTMElection Outcome Translation Matrix
From electoral results to policy probability, with named instrument and calendar bindings.
CBPSCross-Border Payments Stack
SWIFT, CIPS, INSTEX, BRICS Pay, mBridge, and stablecoin rails scored on throughput, sanction durability, and counterparty network.
CABSClimate Adaptation Bond Score
Adaptation-finance instrument grading on additionality, measurability, and sovereign-credit-quality interaction.