The Stock Market Just Did Something for the 2nd Time in 100 Years, and History Says What Comes Next

The article flags that the S&P 500 cyclically adjusted P/E (CAPE) has risen above 40—only the second time in ~100 years (the prior instance was 1999). It argues current earnings could be temporarily overstated because AI-related capex at hyperscalers boosts ecosystem revenue immediately while the costs are recognized over time, potentially making the CAPE look more extreme later. It concludes investors shouldn’t time the market, but cautions against chasing high-multiple growth and instead favor businesses that can perform with or without an AI-driven boom.

Analysis

The market is not pricing a simple bubble risk; it is pricing a belief that AI capex is self-funding and that today’s earnings power is durable. The key vulnerability is a timing mismatch: hyperscalers can boost the ecosystem now while the depreciation and utilization questions hit later, so index-level EPS can look cleaner than the underlying economics for 2-4 quarters. That makes the S&P 500’s multiple more fragile than single-name valuations suggest, especially because a handful of mega-caps dominate both earnings growth and index sentiment.

Near term, the biggest beneficiaries are still the picks-and-shovels names tied to data-center buildout and compute demand, but the asymmetry is shifting from “beat and raise” to “prove ROI.” If capex growth slows even modestly, semis and infrastructure suppliers can de-rate faster than the hyperscalers because their revenue is more directly tied to the spend cycle. Over 6-18 months, the more durable winners are cash-generative, lower-capex compounders like NFLX and quality franchises with less dependence on AI narratives; they can look expensive, but their earnings are less exposed to accounting normalization.

The consensus is missing that this is less about an imminent crash and more about a future margin-air-pocket when the market stops capitalizing growth spend as if it were free. That argues for hedging index concentration rather than making a heroic top call. The thesis is falsified if AI capex keeps accelerating while operating cash flow and free cash flow remain strong through the next two reporting seasons, or if lower rates compress the discount rate enough to offset any earnings-quality reset.

Key Decisions for Investors

• Buy 3-6 month QQQ put spreads on strength as a hedge against multiple compression; best entry is after any failed breakout in mega-cap tech leadership. Risk/reward is attractive if the market starts questioning AI capex ROI before the next earnings cycle.
• Pair trade: long NFLX / short a basket of MSFT, GOOGL, and AMZN for 2-4 months. The long leg is a cleaner cash-flow compounder with less capex overhang; the short leg is where deferred depreciation and capex scrutiny can compress valuation first.
• If wanting express downside exposure, use small short-dated short call risk or put spreads on NVDA only into post-rally strength; do not chase the short here. The thesis needs a catalyst from guidance or capex deceleration, so size modestly and use defined risk.
• Watch the next two hyperscaler earnings prints for capex growth versus operating cash flow. If capex stays elevated but incremental revenue monetization does not improve, rotate out of XLK-style concentration and into quality/value factors.