Research methodology
Where every number comes from.
Authoritative public sources, the current directory dataset, and configured feeds — combined into probability-scored research on AI data center energy risk.
Primary Data Sources
Government & Regulatory
- • Department of Energy (DOE): National energy consumption data, data center electricity projections
- • Lawrence Berkeley National Laboratory (LBNL): Data center energy research and forecasts
- • FERC: Federal energy regulations, Order 2023 interconnection rules
- • State PUCs: Utility tariffs, rate cases, data center-specific regulations
- • EIA: Energy Information Administration statistics and pricing data
Grid Operators (ISOs/RTOs)
- • PJM: Interconnection queue data, capacity auction results, congestion pricing
- • ERCOT: Real-time grid conditions, resource adequacy reports
- • CAISO: Renewable integration data, transmission constraints
- • MISO, SPP, NYISO, ISO-NE: Regional grid data and queue status
Industry & Financial
- • Corporate filings: 10-K/10-Q reports from public data center operators
- • Earnings calls: Management commentary on energy costs and expansion plans
- • M&A databases: Transaction data for infrastructure deals
- • Permit filings: Construction permits and environmental assessments
The GreenCIO Compute Index
The GreenCIO Compute Index is GreenCIO's transparent GPU-hour price benchmark for AI compute — spot and forward compute-price settlements across SKU × region × tenor (H100, H200, B200 and successor accelerators across US-East, US-West, EU, APAC and the Middle East). Every settlement is published as a named series, e.g. GCI-H100-USE-SPOT.
How a settlement is built
- • Source-traced input panel: published venue rate cards and API pricing across hyperscalers, neoclouds, and GPU marketplaces, plus contributed quotes under panel agreements.
- • Filter & weight: outlier trim, freshness decay, and volume-weighting by measured venue capacity — not posted price alone.
- • Settle & band: a trimmed volume-weighted settlement with an explicit uncertainty band and the number of contributing venues.
- • Basis spreads: region-pair, SKU-pair, and neocloud-vs-hyperscaler spreads published alongside the headline settlement.
Transparency & governance
- • Published, versioned rulebook with a rulebook hash on every settlement and a public revisions log.
- • Auditable inputs: hashed input snapshots retained so any historical settlement can be reproduced at the methodology version stamped on the record.
- • Designed to IOSCO Principles for Financial Benchmarks with independent oversight.
- • Administrator independence: the GreenCIO Compute Index is administered independently of any exchange, clearing house, or brokerage, and GreenCIO takes no directional position in any GCI-referenced instrument.
GreenCIO Compute Index benchmarks are available to design partners through the GreenCIO Data API (Pro tier) and the GreenCIO MCP server.
Update Frequency
Configured Feed Cadence
- • News and press releases
- • M&A announcements
- • Regulatory filing alerts
- • GCS-backed feed updates when storage is connected
Source-Dependent Updates
- • Interconnection queue changes
- • Tariff modifications
- • Permit applications
- • Weather and climate data where connected
Weekly Analysis
- • Queue progression analytics
- • Regional trend reports
- • Policy change summaries
- • Market sentiment indicators
Monthly Deep Dives
- • Comprehensive market reports
- • Regulatory landscape updates
- • Infrastructure development tracking
- • ESG metrics compilation
Analytical Framework
Risk Scoring Methodology
Our proprietary risk scores combine multiple factors:
- • Grid Risk (40%): Queue position, interconnection delays, capacity constraints
- • Regulatory Risk (30%): Current tariffs, proposed changes, political climate
- • Energy Cost Risk (20%): Price volatility, renewable availability, demand charges
- • Environmental Risk (10%): Water stress, extreme weather probability, carbon intensity
Predictive Models
Models and rules can be calibrated against historical datasets to forecast:
- • Interconnection approval timelines
- • Energy price trajectories
- • Regulatory change probability
- • Infrastructure build-out patterns
Data Quality Assurance
Validation Process
- Multi-source verification for critical data points
- Automated anomaly detection
- Expert review of outliers
- Historical backtesting
Transparency Standards
- Source attribution for all data
- Confidence intervals provided
- Model assumptions disclosed
- Change logs maintained
Key Differentiators
1. Evidence-First Research: Outputs should preserve source attribution, confidence, and assumptions so analysts can verify before acting.
2. Unified View: Grid, regulatory, financial, and environmental sources can be reviewed in a single risk framework.
3. Forward-Looking Intelligence: Beyond historical data, our models predict future conditions that may impact your investments.
4. Actionable Briefs: Raw data is transformed into research briefs tied to your portfolio and investment strategy where client data is connected.