AI Fear & Greed methodology and sources

Price changes in the TECHi AI coverage basket describe market momentum beyond the Mag-7; they do not measure dollars invested or withdrawn.

Uses market-cap weighted momentum across TECHi-covered AI company tickers when enough providers respond. The full universe is 138 company tickers, ETFs excluded; live coverage is disclosed in the input context.

AI greed is easier to sustain when liquidity is expanding and inflation/rate pressure is not squeezing duration assets.

Blends global liquidity change with every Alpha Vantage macro endpoint wired into the panel: GDP, GDP per capita, Treasury yields, Fed funds, CPI, inflation, retail sales, durable goods, unemployment, and payrolls.

Lower frontier token costs expand the set of AI apps that can work economically, which is a greed signal for the AI stack.

Transforms 1M, 1Y, and AI-era frontier output-token cost changes with ln(baseline/current), caps each horizon, then averages them. Falling cost remains an expansion signal, but large historical deflation cannot pin the component at an outlier.

TECHi's own reader demand: Google Analytics quote-page traffic across covered tickers. A proprietary attention signal, not a market-wide one — capped at a low weight so it cannot dominate the index.

Uses TECHi Google Analytics quote-page demand when available, curated fallback watchlists when analytics is unavailable, and AI ticker concentration. Weight is capped (4/100) because this measures TECHi reader behavior rather than broad market sentiment.

Headline tone checks whether AI coverage is confirming the tape or warning that the market narrative is weakening.

Reads news already stored for 138 covered AI company tickers when a quote page has fetched it. A fresh provider request runs only when that store is empty, with Alpha Vantage NEWS_SENTIMENT as the last fallback. Missing coverage stays neutral.

Ecosystem breadth checks whether AI strength is spreading beyond the headline stocks into frontier models, labor demand, and adjacent risk-technology adoption.

Blends external signals only — frontier-model count, AI job-posting growth, and crypto/risk-technology adoption. TECHi Learn, analyst-desk and AI Power List activity were removed from the score to avoid the index measuring TECHi's own editorial output.

An editorial allocation scenario illustrates concentration across compute, power, labs, apps, tools, and robotics. It is not an observation of money moving between them.

Uses editorial scenario percentages. Higher breadth and lower concentration raise the raw score. Review date is not recorded; these assumptions are excluded from current-input coverage and cannot activate this component.

Technical breadth checks whether the AI universe is participating or only a handful of leaders are holding up the tape.

Scores the 138-ticker AI company universe from quote prints already stored by quote pages. Momentum = 50 + session change percent × 10. Technical score = 70% momentum + 30% of where the price sits between the stored day low and day high. Constructive is a technical score of 60 or higher. Defensive is below 40. The component does not call a provider.

Upcoming AI earnings, macro prints, and weekly delta tape capture the near-term events that can confirm or challenge the current regime.

Scores upcoming earnings density only for TECHi-covered AI company tickers, EPS-estimate coverage, live macro-print availability, and the direction of TECHi weekly delta items.

Gold, oil, gas, bitcoin, inflation, rates, and retail demand are checked as external pressure gauges around AI risk appetite.

Blends the live session moves in Nasdaq 100, NVIDIA, VIX, gold, silver, oil, and bitcoin with the official rates and inflation prints. Equity and bitcoin strength lift the score. A rising VIX or gold price pulls it down.