Latest data: Sep 7, 2026
Data sources

A transparent, free-tier foundation

This MVP uses public and free-tier sources and labels each metric by its real update cadence. Lagging monthly and quarterly series are not presented as daily. Data mode: Supabase.

AI Buildout

Hyperscaler Capex-to-Revenue Divergence

SEC EDGAR XBRL (company filings) • Annual (automated, keyless)

Live updates are computed from each company's PaymentsToAcquirePropertyPlantAndEquipment and revenue XBRL facts via SEC EDGAR's free companyconcept API (no key required), matched by calendar year of fiscal-year-end and only combined when all four companies report both the current and prior year. The single seed value below is one real, precisely computed point (FY2024 to FY2025 combined revenue, from each company's own 10-K/earnings release, +14.7%, versus Epoch AI's reported 2024-to-2025 combined hyperscaler capex growth of +81.4%; note that Epoch AI's capex figure includes Oracle alongside these four, a scope mismatch this one bootstrap point inherits -- the live SEC-only feed replaces it with an apples-to-apples figure once ingestion runs). The four companies have different fiscal year ends (Microsoft's is June 30), so calendar-year bucketing blends slightly offset fiscal periods. Revenue growth is total company revenue, not AI-specific revenue, since AI-only revenue is not separately disclosed.

AI Buildout

Open vs. Closed Model Intelligence Gap

Arena AI code leaderboard (via arena-ai-leaderboards mirror) • Daily (automated, keyless)

This is a watch item, not part of the composite score. Artificial Analysis's own model-comparison API would need a $400/mo Pro plan for this data, so this instead reads github.com/oolong-tea-2026/arena-ai-leaderboards, a free, keyless, daily-scraped mirror of the Arena AI leaderboards that already tags each model's license (open vs. proprietary). That mirror is a single-maintainer side project, not Arena's own feed, and has no durability guarantee -- a prior similar mirror silently stopped updating for over a year with no visible warning. Ingestion checks the mirror's own "latest snapshot" date on every run and refuses to use data more than 4 days old; if that trips, the metric simply stops updating and a GitHub issue is filed automatically on the repo so staleness gets noticed instead of silently going unnoticed.

Credit & Volatility

10Y-2Y Yield Curve

FRED T10Y2Y • Daily

The curve is a macro warning, not a market-timing device. It can stay inverted for a long while before a recession unfolds.

Credit & Volatility

10Y-3M Yield Curve

FRED T10Y3M • Daily

The signal is about macro risk, not crash timing. It can invert well before markets react, and it can normalize before the real economy weakens.

Credit & Volatility

High-Yield Credit Spread

FRED BAMLH0A0HYM2 • Daily

This is a credit market signal, not a direct stock-market forecast. It can stay tight for a long time in a strong economy.

Credit & Volatility

Investment-Grade OAS

FRED BAMLC0A0CM • Daily

Investment-grade spreads can stay tight for long periods in healthy expansions. This is a backdrop signal, not a standalone market forecast.

Credit & Volatility

VIX

FRED VIXCLS • Daily

Low VIX is not a safety signal. It often means the market is underpricing the possibility of a sudden move.

Macro

10Y Treasury Yield

FRED DGS10 • Daily

The yield is not a bubble signal on its own, but it can shape the risk-reward backdrop for the market.

Macro

Sahm Rule

FRED SAHMREALTIME • Monthly

The Sahm Rule is a recession trigger, not an equity signal. It usually confirms stress after unemployment has already begun moving.

Macro

Unemployment Rate

FRED UNRATE • Monthly

It is a lagging indicator, so it is more useful for context than for sounding the first alarm.

Sentiment & Leverage

AAII Sentiment

AAII • Weekly

Sentiment surveys are noisy and can stay extreme for a while. They work best as a secondary check rather than a primary trigger.

Sentiment & Leverage

CBOE Put/Call Ratio

CBOE • Daily

The signal is short-term and can swing on positioning rather than fundamentals. It works best as a sentiment backdrop, not a standalone timing tool.

Sentiment & Leverage

Crypto Speculation

CoinGecko • Daily

Crypto market cap is noisy, young, and structurally different from equities. It should be treated as a sentiment proxy, not a valuation anchor.

Sentiment & Leverage

Margin Debt

FINRA • Monthly

Margin debt trends can be slow and are impacted by regulatory changes and market structure. It is better used as a leverage backdrop than a standalone timing signal.

Valuation

Buffett Indicator

FRED / Wilshire 5000 + GDP • Daily / quarterly

GDP is quarterly and lags the market, so the ratio is smoother and slower-moving than daily price moves.

Valuation

S&P 500 ÷ M2

FRED SP500 / FRED M2SL • Daily / monthly

This is a rough liquidity-based context metric, not a precise valuation tool, and there is no single standard formula for it across analysts.

Valuation

Shiller CAPE

Robert Shiller / Yale • Monthly

The series is a valuation proxy, not a timing tool. Markets can stay expensive for years, especially when growth expectations are strong.

Attribution notes

  • FRED / ALFRED series are used for Treasury, volatility, credit, GDP, labor, and monetary indicators.
  • Robert Shiller / Yale is used for long-run CAPE context.
  • Stooq or Shiller-style long-history index data should be used for S&P history where FRED licensing is too short for dot-com-era charts.
  • CoinGecko is used as the free crypto-market proxy.
  • CBOE and FINRA are listed for the next sentiment and margin-debt additions; they are not silently treated as daily data until implemented.

Back to the dashboard.