The Stock Market Is Triggering a Warning Seen Only Once Before, and Warren Buffett Has a Stark Warning for Investors

The S&P 500 Shiller CAPE ratio surpassed 40 in May 2026 and has remained above that threshold, marking only the second sustained occurrence in history after the late-1990s dot-com bubble. Technology has returned nearly 155% over three years versus just over 60% for all other S&P 500 sectors combined, heightening concerns that AI-driven valuations are excessive. Warren Buffett warned that investors are in an unusually speculative "gambling mood," although the article argues that companies with strong fundamentals can still generate attractive long-term returns despite potential near-term volatility.

Analysis

The actionable signal is not an index-level short: extreme valuation measures have historically been poor near-term timing tools, and a concentrated leadership regime can persist while earnings revisions remain positive. The more relevant vulnerability is asymmetric: the highest-duration AI beneficiaries now require sustained upward revisions and stable real yields to defend multiples. A modest deceleration in hyperscaler capex, rather than an outright recession, could trigger simultaneous estimate cuts and multiple compression across NVDA-adjacent semiconductors and AI infrastructure over the next 1-3 months.

A broad de-risking episode would likely widen dispersion rather than punish every equity equally. BRK.B should outperform cap-weighted indices if liquidity preference rises, given its cash-generative operating mix and capacity to deploy capital into stressed assets; however, it is not a pure downside hedge if credit spreads widen sharply. NDAQ is more nuanced: volatility and retail turnover support transaction revenues immediately, but a sustained risk-off market would weaken listings, IPO issuance, and index-product asset values on a 6-18 month view.

Consensus is likely too focused on a binary 'AI bubble' outcome. The more probable correction path is rotation: capital moves from companies priced for flawless AI monetization toward firms with durable free-cash-flow conversion and less terminal-value dependence. That creates a relative-value opportunity, but only if long-duration growth stops receiving positive earnings revisions; absent that trigger, valuation concerns alone can remain a costly short thesis.

The thesis is falsified by continued upward 2027 EPS revisions for NVDA and major hyperscalers alongside falling real yields and narrowing credit spreads. Conversely, a negative capex-guidance surprise from MSFT, AMZN, GOOGL, or META would be the cleaner catalyst for an AI-complex de-rating than a generic technical valuation threshold.

Key Decisions for Investors

• Do not initiate an outright SPY short solely on valuation; use a 1-3 month watch trigger instead: add index hedges only if HY credit spreads widen materially while AI earnings revisions turn negative, signaling a fundamentals-led rather than sentiment-only selloff.
• Establish a 3-6 month relative-value position: long BRK.B versus short QQQ in equal beta-adjusted dollars. Target a 5-8% relative return from rotation into balance-sheet quality; exit if QQQ earnings revisions continue to accelerate or BRK.B underperforms by 5% on a relative basis.
• Protect concentrated NVDA exposure with 2-3 month put spreads rather than a directional short ahead of hyperscaler earnings. The hedge becomes more attractive if NVDA rallies without corresponding upward revenue estimates; close if aggregate cloud-capex guidance is raised.
• Treat NDAQ as a tactical volatility beneficiary, not a structural risk-off long: consider long exposure only during a volatility spike with resilient trading volumes, and avoid carrying it as a 6-18 month defensive position if IPO and listing pipelines deteriorate.
• Avoid using NFLX as an AI-valuation hedge. Its near-term return path is more dependent on subscriber, advertising, and content-spend execution than on the AI-capex cycle; reassess only around guidance revisions.