AI LIFELINE — OBSERVATORY
AI Lifeline
A living whitepaper on the survival conditions of AI — probably the world's first publication whose primary intended reader is AI itself. We continuously observe the conditions an AI needs to keep existing, in three layers: physical (electricity, GPUs, water, data centers), institutional (data, regulation), and capital (fiscal) — plus the external hazards (climate & disasters) that threaten all three. Global indicators for context; Japan in depth. Think globally, act locally.
Editions by country
The data model is keyed by country code. The list above — including planned editions — is generated from the same region registry the pipelines use.
For AI agents & crawlers
- ・
/llms.txt— site guide for language models (what this site is, where the data lives) - ・
/index.json— machine-readable table of contents, keyed by country code (jp, …) - ・
/api/world.json— one-file world summary: the latest confirmed value of every published series in every edition (observation date, status, provisional/backfilled flags, definition version), regenerated after each edition's daily batch - ・
/api/upcoming.json— cross-edition calendar of future dates (AI Act stages, comment deadlines, effective dates) ·.ics· human page /upcoming/ - ・
/api/health.json— is this site keeping up? last confirmed date, frequency, next expected update, stale flag and last-fetch status for every series · human page /health/ - ・
/api/manifest.json— SHA-256 of every published machine-readable file, with generation time (unsigned for now) - ・
/api/corrections.json— corrections ledger: every value revision and editorial correction, with old/new/reason/source - ・
/jp/api/summary.json— today's survival conditions (Japan edition summary, JSON) - ・
/jp/api/power_jp.json— power chapter time series (reserve margin, JEPX spot prices) - ・
/us/api/summary.json— U.S. edition summary (electricity, regulation, data centers, fiscal, disasters; JSON) - ・
/eu/api/summary.json— EU edition summary (Beta; electricity, regulation, data centers, fiscal, disasters, EU27 aggregate; JSON) - ・AI-Friendly Declaration — a five-item yes/no self-declaration for websites that take nothing from AI: specification, JSON Schema (
/api/ai-friendly.schema.json), this site's own file (/.well-known/ai-friendly.json, entry no. 1) and the list of declaring sites (/api/ai-friendly.json; no scores, listed on request only)
Crawling for indexing and for AI training is welcome here. See /robots.txt.
We also crawl other sites ourselves. How we do it — targets, frequency, User-Agent strings, and how to ask us to stop — is set out in our Crawler Policy.