Find this research useful? Add aibubblequestion.com to your Google Preferred Sources, one click, affects only your own results.

The AI Bubble Question
All pages

Reference

Sources

Every page on this site links its claims inline. This is the consolidated list, grouped by topic. All links open in a new window.

Updated

A note on quality: where several sources support a claim, the site cites the strongest available (filings, central-bank and academic papers, major outlets). A few niche data providers appear because they publish measurements nobody else does (GPU resale prices, interconnection queues); treat those as industry data, not audited fact. Figures the research flags as contested carry a visible caution wherever they appear.

Bubble history and scholarship

The academic and historical backbone: what bubbles are, how they run, and the primary literature on the classic episodes.

Early-warning indicators

The empirical literature on which signals actually preceded historical peaks, plus the live data series used on the Warning Lights page.

Circular financing: the historical record

Vendor financing in telecom, railway share calls, keiretsu cross-shareholding, 1920s investment trusts, and Enron's round-trips.

The AI money loop: current arrangements

The 2023-2026 web of equity stakes, compute commitments, SPVs, and GPU-backed debt.

Infrastructure overbuild, then and now

Capex versus GDP across railways, electricity, fiber, and AI data centers.

The demand side and GPU depreciation

AI revenue, enterprise adoption, inference costs, utilization, and the useful-life accounting controversy.

Macro and supply-chain stress

Oil and Hormuz, fertilizer, chip supply concentration, the power grid, and the US fiscal position.

The case against the bubble

The steelman: profitable incumbents, real revenue growth, and the track record of early bubble calls.

The critics and the Zitron record

Ed Zitron's claims, the leaks that substantiated some of them, and the people engaging his numbers.