How It Works
Transparency in action — Learn how we collect, validate, and publish AI environmental data from public sources to create the most comprehensive sustainability ranking.
Our Mission
GreenCodeAI is an activist transparency platform — the 'Greenpeace of AI' — that exposes the real environmental impact of artificial intelligence. We believe that transparency drives change, and that every consumer, investor, and journalist should have access to verified data about AI companies' environmental footprint.
100% Independent
No corporate sponsorships from ranked companies. No conflicts of interest. We accept no funding from AI vendors or cloud providers.
Open Source
All our code, algorithms, and data sources will be public at launch (Q1 2026). Anyone will be able to audit, contribute, or replicate our work.
Data Integrity
We only use verifiable public data: ESG reports, scientific papers, government databases, and verified APIs. No estimates without clear disclaimers.
How We Collect Data
A fully automated and transparent pipeline
Our data collection process combines multiple sources to create the most comprehensive view of each AI company's environmental impact. Everything is automated, scheduled, and version-controlled.
Our Data Sources
1. Corporate ESG Reports
We scrape sustainability reports published by companies (Google, Microsoft, Meta, Amazon, etc.)
Examples: Google Environmental Report, Microsoft Sustainability Report, Meta Sustainability Report
Updated: Annually (when companies publish new reports)2. Public APIs
We integrate real-time data from verified environmental APIs
- •Electricity Maps API — Real-time grid carbon intensity by region (gCO₂/kWh)
- •Google Cloud Carbon Footprint — Public carbon data for GCP datacenters
- •ML CO2 Impact — Estimated emissions for AI models based on scientific research
- •Company APIs — Some companies (like Anthropic, Hugging Face) provide public metrics
3. Scientific Research
Peer-reviewed studies and academic papers about AI energy consumption
- •"Energy and Policy Considerations for Deep Learning in NLP" (Strubell et al., 2019)
- •"Carbon Emissions and Large Neural Network Training" (Patterson et al., 2021)
- •University research labs (Stanford HAI, MIT Climate, Berkeley AI Research)
4. Manual Verification
When data is incomplete, our team manually verifies through:
- •Press releases and company blog posts
- •Investor presentations (ESG sections)
- •Regulatory filings (CSRD compliance in EU)
- •Datacenter specifications (PUE, WUE, renewable %)
The Data Pipeline
From raw data to rankings
Our automated ETL (Extract, Transform, Load) pipeline runs on a weekly schedule, ensuring our rankings reflect the latest available data.
Scraping & Extraction
Every Monday at 3:00 AM UTC, our Python scraping bots extract data from corporate reports, press releases, and research papers.
Tech: Scrapy (web scraping) + Playwright (JavaScript-heavy pages) + Pydantic (data validation)
API Integration
We fetch real-time data from Electricity Maps, Google Cloud Carbon Footprint, and other verified APIs. Carbon intensity data is updated daily.
Tech: REST APIs + rate limiting + caching (5-minute intervals)
Data Validation
All collected data passes through validation rules: detecting outliers, cross-referencing sources, flagging missing data.
- ✓ Cross-reference with 2+ sources when possible
- ✓ Flag values outside expected ranges (outliers)
- ✓ Verify datacenter locations via geolocation APIs
- ✓ Timestamp all data points for historical tracking
Score Calculation
We apply our open methodology to calculate the Environmental Impact Score (EIS) for each company. All algorithms are public on GitHub.
See detailed methodology →Storage & Versioning
Processed data is stored in PostgreSQL (relational data) with full version history. Every change is tracked with timestamps.
Tech: Supabase PostgreSQL + Time-series optimization
Publication
Updated rankings are automatically published to the website. Our public API makes all data available to journalists, researchers, and developers.
Tech: Next.js ISR (Incremental Static Regeneration) + Public REST API
Update Schedule
Our data is refreshed automatically:
Daily
Carbon intensity updates (Electricity Maps API)
Weekly
Web scraping of news, blog posts, press releases
Monthly
Full ranking recalculation with new data
Quarterly
ESG reports ingestion + methodology review
Radical Transparency
Every step is auditable
Unlike proprietary ranking systems, we make everything public. Our commitment to transparency is non-negotiable.
Open Data
All our processed data will be available via:
- •Public API (free, rate-limited to 100 req/hour)
- •CSV/JSON exports (download raw rankings)
- •Open datasets (historical data available at launch)
Open Algorithms
Our scoring methodology is fully documented:
- •Methodology page — Explains EIS formula in detail
- •Source code — All calculations will be public at launch (Q1 2026)
- •Peer review — We welcome scientific feedback
Community Corrections
We welcome corrections from anyone:
- •Contact form — Report data errors or outdated info
- •Submit Data form — Companies can submit official metrics
- •Community contributions — Open at launch (Q1 2026)
No Conflicts of Interest
We maintain complete independence:
- •❌ No funding from ranked AI companies
- •❌ No cloud provider sponsorships
- •❌ No pay-to-improve or pay-to-remove options
- •✓ Funded by donations, ESG-aligned partners, paid audits
Our Impact
How transparency drives change
By exposing environmental data, we empower multiple audiences to make informed decisions and pressure companies to improve.
Consumers & Developers
Choose greener AI providers
- →Prefer companies with high EIS scores
- →Avoid blacklisted companies with poor transparency
- →Support open-source models hosted on renewable energy
ESG Investors
Assess sustainability risks
- →Integrate EIS scores into investment decisions
- →Demand better environmental reporting from portfolio companies
- →Benchmark companies against industry averages
Journalists & Media
Investigate greenwashing
- →Expose companies making false sustainability claims
- →Compare public commitments vs. real metrics
- →Use our data in investigative journalism
Policymakers & Regulators
Inform AI sustainability policies
- →Reference our data in climate legislation
- →Mandate environmental reporting requirements
- →Establish industry benchmarks for AI emissions
Green List & Blacklist
We amplify the best and expose the worst:
Green List
Companies that excel in transparency and sustainability
Criteria: 100% renewable energy, PUE <1.2, quarterly public reporting, verified certifications
Blacklist
Companies with poor transparency or high environmental impact
Criteria: No public data >12 months, high-carbon datacenters without offsets, water stress violations
Join the Movement
Every AI request has a carbon cost. It's time to make that cost visible.
💚 Code will be open source at public launch (Q1 2026)