Warren Buffett Just Issued a Blunt Message as This Stock Market Warning Rings for the Second Time in 155 Years

The S&P 500 Shiller CAPE ratio stands near 41, versus its 155-year average of 17.8 and close to the dot-com peak of about 44, signaling elevated equity valuations and potentially more modest medium-term returns. Warren Buffett has warned that speculative behavior is making value opportunities difficult, although the article argues that today's AI leaders have substantial cash flow and operating businesses unlike many dot-com-era companies. The recommended approach is to remain invested in durable, moat-protected businesses rather than attempting to time a potential market correction.

Analysis

The relevant risk is not an index-level collapse but a dispersion event: expensive benchmark exposure has become increasingly dependent on a small set of AI-linked companies sustaining elevated earnings revisions. If long-duration rates rise or AI monetization timelines slip, NVDA, AVGO and TSM face the largest multiple-and-cycle risk because their valuation support depends on continued accelerator demand and customer capex. In contrast, GOOG, META, MSFT and AMZN have offsetting cash-flow streams, but their returns will increasingly be judged on incremental AI revenue and margin conversion rather than infrastructure spending.

A high CAPE is a poor 1-3 month timing tool, but it is useful for setting a lower expected-return hurdle over 6-18 months. The non-obvious implication is that passive SPX exposure offers less diversification than headline constituent count suggests: an AI capex disappointment would simultaneously hit semiconductor revenue, hyperscaler depreciation, data-center power demand and index multiples. BRK.A/B is comparatively insulated through its cash generation, operating-business exposure and capacity to deploy liquidity during a risk-off episode; it is a more credible quality/optionality vehicle than simply rotating into crowded mega-cap defensives.

Consensus may be too focused on whether AI is a bubble. The investable distinction is between a demand pause and a return-on-invested-capital failure: a temporary digestion period can hurt chip earnings sharply while improving hyperscaler free cash flow, whereas weak enterprise adoption would impair both groups. Track quarterly capex guidance, cloud growth and disclosed AI revenue/booking metrics; absent a downward revision in those indicators, broad de-risking is premature despite rich valuations.

Key Decisions for Investors

• Reduce concentrated NVDA/AVGO/TSM beta into strength; retain exposure only through a 1-3 month pair trade long GOOG or META versus short SMH, targeting a 10-15% semiconductor relative drawdown if 2027 capex guidance moderates. Exit if aggregate hyperscaler capex guidance is raised by more than 10%.
• Add BRK.B as a 6-18 month quality-and-optionality allocation versus SPY, sized as a hedge against valuation compression rather than a crash trade. Thesis is falsified by sustained underwriting deterioration, a material decline in operating cash generation, or a relative-performance break below the prior 12-month range.
• For existing MSFT and AMZN longs, demand evidence before adding: watch cloud growth acceleration alongside stable free-cash-flow margins in the next two earnings cycles. If capex rises while cloud growth and AI monetization fail to accelerate, rotate exposure toward GOOG/META, where ad-funded cash flow provides greater near-term funding capacity.
• Do not initiate an outright SPY short solely on valuation. Set a risk alert if real yields rise 50 bps, AI leaders cut forward revenue expectations, or credit spreads widen materially; the combination would justify replacing index beta with cash and BRK.B rather than attempting to time an isolated valuation signal.