The atlas · market signals
Find the report before it's obvious.
Bezos didn't find Amazon by reading trend pieces about e-commerce. He found one number: internet usage growing 2,300% a year. This page is that move, generalized — real category-level adoption and production curves, not mention counts. /trends already covers attention (what's being talked about) and explicitly treats that as crowding, not demand. This is the other side: what the world is actually using, building, or spending on.
2.5 → 111.5 over 29 years (+4361% total). Turns out this is NOT the saturated comparison case it was expected to be — mobile compounded faster annually than every other category here, over a longer window than any of them. One of history’s fastest technology diffusions, still showing up in the numbers three decades later.
| # | category | baseline | latest | window | total growth | per year |
|---|---|---|---|---|---|---|
| 01 | Mobile subscriptionsTurns out this is NOT the saturated comparison case it was expected to be — mobile compounded faster annually than every other category here, over a longer window than any of them. One of history’s fastest technology diffusions, still showing up in the numbers three decades later. | 2.5 (1996) | 111.5 (2025) | 29y | +4361% | +14% |
| 02 | Fixed broadbandSubscriptions per 100 people — infrastructure capacity, not just access. A leading indicator for anything that needs real bandwidth (video, cloud gaming, remote work tooling). | 3.4 (2005) | 20.1 (2025) | 20y | +491% | +9% |
| 03 | Internet adoptionThe literal Bezos-era metric, run forward 30 years. Growth has naturally slowed as it approaches saturation — that deceleration is itself a real signal. | 15.6 (2005) | 73.6 (2025) | 20y | +372% | +8% |
| 04 | High-technology exportsShare of manufactured exports that are high-tech — a shift metric, not a volume metric. Rising share means the world is manufacturing more complex goods, not necessarily more goods overall. | 19.9 (2007) | 24.7 (2024) | 17y | +25% | +1% |
| 05 | R&D expenditureShare of GDP spent on research and development — the closest thing World Bank tracks to "how much is the world investing in finding the next thing." | 2.0 (1999) | 2.6 (2023) | 24y | +27% | +1% |
| 06 | Renewable energy consumptionShare of final energy consumption from renewables — directly relevant to the energy-abundance problem already ranked on this index. | 17.4 (1996) | 19.7 (2020) | 24y | +13% | +1% |
source: World Bank Open Data · indicators verified live 2026-08-07
How this is built, and what's next
Usage, not attention. Every series here is World Bank Open Data — real adoption, production, or spend, worldwide, keyless and free. Nothing here is a mention count, a search-interest score, or a headline tally.
Ranked by annualized growth, not total growth. Each indicator has its own history and reporting lag — mobile and renewable-energy data go back to 1996, internet and broadband only to 2005, high-tech exports only to 2007 — so a raw baseline-to-latest percentage isn't comparable across rows: a 29-year window will out-accumulate a 17-year one regardless of which is actually compounding faster. Total growth is still shown for context; the ranking uses CAGR (the standard fix for comparing unequal windows).
Six categories today. More need a free key. FRED (840,000+ US economic series — GDP by industry, consumer spend by category) and UN Comtrade (import/export volume by product) would add real firepower, but both now require a free API key — a 2-minute signup, same pattern as this project's other credentials. Neither is wired up yet.