01HYPERSCALER AI CAPEX
Quarterly & trailing-12-month capital expenditure, Microsoft / Alphabet / Amazon / Meta.
Combined quarterly capex ($B)
Capex by company ($B, quarterly)
Latest quarter detail
Fields: capex_quarterly, capex_ttm, capex_yoy_growth. Source: company 10-Q / 10-K cash flow statements.
02CAPEX VS CASH GENERATION
Operating cash flow, free cash flow, and the capex/OCF ratio — flags where capex is outrunning cash generation.
Capex growth vs. FCF growth (YoY %)
Capex as % of operating cash flow
Cash generation detail — latest quarter
Fields: operating_cash_flow, capex, free_cash_flow, fcf_margin, capex_pct_ocf. Source: 10-Q cash flow statements (calculated: FCF = OCF − Capex).
03AI MONETIZATION
Disclosed AI-attributable revenue signals. Reported vs. analyst-estimated figures are labeled explicitly.
Most hyperscalers do not separately break out "AI revenue" in filings — figures below are either a disclosed growth-rate proxy (e.g. Azure AI contribution to Azure growth) or third-party/analyst estimates, labeled accordingly. Cells with no reliable source read "Not disclosed."
AI revenue growth (YoY %, where disclosed)
Combined AI-linked capex ($B, TTM)
Est. AI revenue ÷ AI infra capex (TTM)
04NVIDIA
Revenue, Data Center segment, margins, inventory and receivables — the clearest read on hyperscaler demand pass-through.
Revenue vs. Data Center revenue ($B)
Gross & operating margin (%)
Inventory & receivables growth vs. revenue growth (YoY %)
Signal detail
Fields: revenue, datacenter_revenue, gross_margin, operating_margin, inventory, receivables, yoy_growth. Source: NVDA 10-Q / 10-K, earnings press release segment tables.
05MEMORY / HBM
Micron as primary disclosed source; SK Hynix / Samsung shown where reliable public data exists.
Micron revenue & HBM/data-center revenue ($B)
Micron gross margin & inventory days
Fields: revenue, hbm_revenue, gross_margin, inventory, dram_price_index. Source: Micron 10-Q/10-K; SK Hynix & Samsung figures limited to disclosed segment data where available, else "Not disclosed."
06DEPRECIATION WALL
The growing depreciation burden from recent AI infrastructure builds — flagged when D&A outgrows revenue/operating income.
Combined capex vs. combined depreciation ($B)
D&A growth vs. revenue growth (YoY %)
Fields: ppe_gross, capex, depreciation_amortization. Source: 10-K PP&E footnotes and cash flow statements (calculated: capex − D&A = net investment gap).
07ELECTRICITY / DATA-CENTER DEMAND
Announced vs. actual operating capacity — the physical-world constraint on AI infrastructure buildout.
US data-center electricity demand (TWh/yr)
Announced vs. actual operating capacity (GW)
Fields: electricity_actual_twh, electricity_projected_twh, announced_capacity_gw, operating_capacity_gw, grid_requests_gw, cancelled_projects_gw. Source: EIA, DOE LBNL data-center reports, grid operator interconnection queues.
08S&P 500 CONCENTRATION
Combined index weight of the core AI-capex complex — the transmission channel to the broader market.
Combined weight over time (%)
Current weight by company (%)
Fields: index_weight_pct per company. Source: S&P index committee published weights / index provider fact sheets.
09AI SUPPLY CHAIN MAP
How a hyperscaler capex slowdown would propagate backward through the chain.
10WARNING SIGNALS
Composite read across every tracked indicator, plus the core spending-vs-return relationship.
DATASCHEMA, SOURCES & LIVE-DATA CONFIGURATION
1. Standardized record schema
Every metric in this dashboard — regardless of ultimate source (CSV, JSON, API, SEC filing, FRED) — is expected in this flat shape. The rendering layer reads only this shape, so replacing DATA_SOURCE below is the only change required to go live.
{
"company": "MSFT", // ticker or entity id
"ticker": "MSFT",
"cik": "0000789019",
"quarter": "Q3",
"fiscal_year": 2025,
"period_start": "2025-01-01", // null for instant (balance-sheet) facts
"period_end": "2025-03-31",
"metric": "capex_quarterly", // see field list below
"value": 19500000000, // null when not_disclosed
"unit": "USD",
"source": "SEC EDGAR XBRL Company Facts",
"api_source_url": "https://data.sec.gov/api/xbrl/companyfacts/CIK0000789019.json",
"filing_source_url": "https://www.sec.gov/Archives/edgar/data/789019/000...-index.htm",
"accession_number": "0000789019-25-000045",
"form": "10-Q",
"xbrl_concept": "PaymentsToAcquirePropertyPlantAndEquipment",
"reported_or_estimated": "reported", // reported | calculated | not_disclosed
"last_updated": "2025-10-24"
}
A record is reported when it's a single disclosed XBRL fact, calculated when this pipeline derived it (a standalone quarter from cumulative facts, or a ratio like FCF margin), or not_disclosed when SEC does not standardize the metric at all (value is always null in that case — never estimated).
2. Required external data fields
3. Recommended data source per field group
4. Bubble Risk Index — calculation transparency
The index is a weighted composite of five sub-scores, each normalized 0–100 from underlying growth-rate gaps (never fabricated — see computeBubbleIndex() in the page source). Weights are configurable in CONFIG.riskWeights.
5. Instructions — replacing demonstration data with live data
1
Keep the schema. Produce a JSON array of records matching the schema in step 1 — one record per company/quarter/metric.
2
Point CONFIG at it. In the page source, set
CONFIG.dataSource.mode to "json" or "api" and set CONFIG.dataSource.url (or path for a local file).3
No other code changes needed.
loadData() fetches the records, reshapeData() pivots them into the per-section structures the charts already read, and every chart, table, and the risk index recompute automatically.4
Missing values. Omit a record rather than inventing one — the renderer displays "Not disclosed" for any metric it can't find for a given period.
5
Provenance stays visible. Keep
source, source_url, and reported_or_estimated populated on every record — these render as the footnotes and badges throughout the dashboard.6. Live-data connection config (edit in page source, top of <script>)
const CONFIG = {
dataSource: {
mode: "demo", // "demo" | "json" | "api"
path: "./data/ai_bubble_dataset.json", // used when mode === "json"
url: "https://your-api.example.com/v1/ai-bubble-metrics", // used when mode === "api"
apiKeyHeader: null // e.g. "X-API-Key" — set if your API requires auth
},
riskWeights: {
capexAcceleration: 0.20,
capexFcfGap: 0.25,
depreciationGap: 0.20,
nvidiaBalanceSheet: 0.20,
indexConcentration: 0.15
}
};