Source principles
TheAIMeters prioritizes public, verifiable and reproducible sources whenever possible. Some indicators are direct snapshots from public APIs or public pages, while others are estimates derived from public signals and documented assumptions.
Types of sources used
Public platform data
Some indicators use public counts from platforms such as Hugging Face, GitHub, arXiv, Stack Overflow, OpenAI status pages and job listing APIs.
Research and technical literature
Environmental and infrastructure estimates rely on public research, technical reports, energy assumptions and published analysis about data centers, GPUs and AI workloads.
Infrastructure assumptions
For estimated counters, TheAIMeters applies explicit assumptions about workload growth, GPU usage, electricity demand, water intensity and carbon intensity.
Status and reliability feeds
Operational indicators may use public service status feeds and rolling time windows to summarize visible service incidents.
Reference sources
Hugging Face
Used for public model and dataset ecosystem indicators.
Visit sourcearXiv
Used as a public signal for AI research publication activity.
Visit sourceGitHub
Used as a public signal for machine-learning and AI software repository activity.
Visit sourceStack Exchange API
Used for public Stack Overflow tag and question metadata.
Visit sourceOpenAI Status
Used for publicly reported OpenAI service incident indicators.
Visit sourceAdzuna API
Used for public AI job listing estimates across selected markets.
Visit sourceIEA
Used as a reference for energy, data center and electricity-related context.
Visit sourceLimitations
Not all AI activity is publicly disclosed. Many indicators should be interpreted as directional estimates or public snapshots rather than audited measurements. Source availability, definitions and reporting methods may change over time.
Corrections and feedback
If you notice an outdated source, a better reference or a methodological issue, you can contact us at contact@theaimeters.com .