FUTURE3.AI · CASSANDRA
AI-Cycle Risk Monitor · as of 2026-08-21
Conditions monitor · not a forecast

Cassandra

Is the AI build-out overheating? A transparent, real-data read of the AI / semiconductor cycle across value, fundamentals, supply chain and the macro / Fed backdrop — a composite, not a single trend line. High = stretched and fragile.

CALMHIGH
50.8
elevated
6 of 6 panels live · weights valuation 22% · fundamentals 18% · supplychain 14% · macro 22% · crowding 12% · adoption 12%
CAPE 41.79 — 99.1th percentile (richer only at the 2000 peak)Market breadth at an extreme — equal-weight at the 3.2th percentile (narrowest in years)NVDA = 44.9% of the semis complex — single-name dominanceSemis fundamentals strong — PEG 0.91, margins +10.7pp (cushions valuation)Chinese models = 60.7% of OpenRouter token volume · 41.6% of calls
01 / VALUATION · BROAD MARKET
CAPE & risk premium
94.7/100
Cyclically-adjusted earnings — immune to single-year EPS, ranked vs 155 years. A negative risk premium means equities yield less than Treasuries.
Shiller CAPE41.79 · 99.1ᵖᶜᵗ
Equity risk premiumnullpp vs 10y null%
02 / FUNDAMENTALS · SEMICONDUCTORS
Earnings quality
7.6/100
The counterweight: do profits justify the price? PEG<1 = growth more than pays for the multiple; expanding margins cushion rich valuations. Low score = healthy.
Median PEG0.91
Operating margin Δ+10.7pp
Earnings growth160% YoY · ROE 40%
namefwd P/EPEGΔmgn
NVDA16.70.6+16.5
AMD30.31+19
AVGO18.70.41+9.5
TSM19.10.98+10.7
ASML292.1+2.4
SMCI6.90.91+2.9
MU6.30.14+57
03 / SUPPLY CHAIN · BUILD-OUT
Capex & GPU rents
15.6/100
MSFT · GOOGL · AMZN · META capex (the AI build-out) and H100 rental prices. A parabolic ramp builds capacity ahead of demand; depressed rents signal oversupply.
Hyperscaler capex$116B · -67.6% YoY
H100 rent$2.27/gpu-hr
04 / MACRO & FED
Liquidity & risk appetite
65.9/100
Is the cycle being fuelled or squeezed? Low VIX & tight spreads = complacency; an inverted curve and tightening conditions are what end build-outs. FRED + CBOE, decades deep.
VIX15.5 · 34.1ᵖᶜᵗ
HY credit spreadnull% · nullᵖᶜᵗ
Yield curve 10y−2y
Financial conditions
05 / CROWDING & POSITIONING
Breadth & concentration
97.7/100
How narrow is the trade? A low equal-weight / cap-weight ratio means a handful of mega-caps hold up the index — thin breadth, the "who's left to buy" fragility. Orthogonal to VIX.
Equal-wt / cap-wt0.2889 · 3.2ᵖᶜᵗ
Semis vs market0.685 · 99ᵖᶜᵗ
NVDA share of complex44.9%
06 / AI DEMAND · CHINA vs US
OpenRouter call share
1.7/100
Real inference traffic through OpenRouter — Chinese-origin models (DeepSeek, Qwen, GLM, Kimi…) vs US/West (GPT, Claude, Gemini…). China 60.7% of token volume · 41.6% of calls · 2026-08-20.
China 60.7% tokensUS/West 39.3%
deepseek/deepseek-v4-flash-2021825.5B
tencent/hy3-202607061254.5B
xiaomi/mimo-v2.5-202604221228.0B
deepseek/deepseek-v4-flash-202861.3B
nvidia/nemotron-3-ultra-550b-a739.1B
TRACK RECORD · TODAY vs HISTORY
Where today sits
Today vs decades of market history
CAPE (Shiller, 1871→)41.79 · 99.1ᵖᶜᵗ
VIX15.5 · 34.1ᵖᶜᵗ
Equal-wt / cap-wt (23y)0.2889 · 3.2ᵖᶜᵗ
Semis vs market0.685 · 99ᵖᶜᵗ
Cassandra’s own daily record
034671002026-06-232026-08-21

Tracked since 2026-06-23 · 60 days · today 50.8 (elevated) · range 46.8–79.2 · mid-range

categorytodaytrajectoryΔ since start
Valuation94.7+3.9
Fundamentals7.6-7.4
Supply chain15.6-56.0
Macro & Fed65.9+14.2
Crowding97.7-1.3
AI demand1.7+1.7

What this is — and what it isn't

Cassandra is a transparent conditions monitor, not a crash predictor. It blends six real-data categories — valuation (Shiller CAPE & the Fed-model risk premium), fundamentals (semis margins, growth, PEG), supply chain (hyperscaler capex & GPU rents), macro / Fed (VIX, credit spreads, the yield curve, financial conditions), crowding (equal- vs cap-weight breadth, semis concentration) and AI demand — into one gauge. Each sub-score is ranked against real history or scored by a transparent, documented rule; weights are a judgement (valuation 22% · fundamentals 18% · supplychain 14% · macro 22% · crowding 12% · adoption 12%), not fitted. A category without enough data is shown as "building history" and excluded — never given a fabricated score.

AI-demand caveat: the China-vs-US panel is the OpenRouter daily call/token share — real inference traffic, by model origin (Chinese labs vs US/West), through one major router. It is a true usage cross-section but not all inference: it excludes first-party API traffic (direct OpenAI, Anthropic, Google) and self-hosted deployments. CAPE is a broad-market GAAP measure; the semis fundamentals are current snapshots with a multi-quarter margin trend.