The Stock Market Is Doing Something for Only the 3rd Time in 156 Years, and History Is Clear a Crash May be Hiding in Plain Sight

The S&P 500 CAPE ratio is around 41 as of October 2026, versus a historical average near 17.8, while the index has gained about 115% since its October 2022 low. The article warns that elevated valuations and reliance on continued AI infrastructure spending could leave less room for disappointment, but notes that today’s AI leaders generate earnings and that a high CAPE does not time a correction. Morgan Stanley projects capex at five major hyperscalers of nearly $800 billion in 2026, $1.2 trillion in 2027, and $1.4 trillion in 2028; the author suggests trimming concentrated AI holdings rather than selling everything.

Analysis

CAPE is a warning about prospective long-run returns, not a reliable near-term sell signal; its backward-looking earnings base may also understate how much current AI profits have changed the index mix. The more actionable risk is a capex-to-returns mismatch: if hyperscalers slow spending or fail to show monetization, pressure can travel quickly from cloud budgets to GPU orders, HBM demand, and supplier expectations. That makes NVDA and MU exposed not only to valuation compression but also to a reversal in customer spending and the risk that capacity arrives ahead of demand. Conversely, a capex pause could preserve cash for buybacks and dividends at AMZN, META, MSFT, GOOG, and ORCL, partly cushioning their equity downside even as supplier multiples re-rate.

Over the next 1–3 months, track hyperscaler earnings commentary on capex, utilization, and AI revenue conversion—not announced budgets alone. Over 6–18 months, power availability, grid equipment lead times, and regulated cost recovery may determine which infrastructure beneficiaries actually earn returns; utilities and electrical-equipment firms are not automatically winners if financing costs or permitting delay projects. The contrarian point: expensive aggregate valuation does not establish that AI leaders are a bubble, but profitable growth does not immunize them from a capex cycle. The thesis weakens if capex converts into sustained incremental cloud/AI revenue and supplier orders without weakening returns on capital; it strengthens on budget cuts, lower utilization, or guidance/order revisions.

Key Decisions for Investors

• For portfolios materially overweight AI semiconductors, trim exposure in stages rather than liquidating on the CAPE signal. Reassess after the next AMZN, GOOG, META, MSFT, and ORCL updates; require evidence of monetization and utilization to restore risk.
• Avoid a broad short based solely on valuation. As a conditional relative-value expression, consider reducing NVDA/MU versus diversified power and electrical-infrastructure exposure, subject to checking current valuation, order backlog, and project economics; the article supplies none of those company-level inputs.
• Set an alert for hyperscaler capex guidance and AI/cloud revenue conversion. A spending slowdown or weaker supplier-order outlook is a catalyst to cut semiconductor risk; stable or rising monetization alongside sustained orders would falsify the bearish capex-cycle thesis.