The Stock Market Is About To Do Something It's Never Done This Century. History Says This Is What Happens Next (Hint: It's Bullish)

FactSet estimates S&P 500 EPS will rise 50.4% in Q2—the fastest pace since Q2 2021 (91.6%), when EPS growth was in a rebound from the pandemic. The article argues the surge is supported by real AI infrastructure build-out (e.g., Micron GAAP net income of $28.2B, up from $1.89B y/y; cited as ~4% of Q2 index growth) rather than a pure earnings decline rebound. Despite bubble/valuation concerns via a near-peak CAPE reading, it notes forward P/E around 20 as a justification for sustaining AI-driven profit growth priced into valuations.

Analysis

The real signal is not that AI is boosting earnings; it is that the profit pool is becoming more concentrated upstream. That favors the compute and memory vendors with pricing power and production bottlenecks, while leaving broad-market multiples more vulnerable because a growing share of index EPS is being driven by a handful of capital-intensive winners and some low-quality accounting marks. In other words, the market can be right on earnings and still be fragile on breadth.

The second-order risk is that investors are extrapolating current AI capex as if it were a straight-line demand curve. If hyperscaler spend pauses for even one quarter, the most levered beneficiaries will be the first to gap down because their earnings are tied to utilization, mix, and memory pricing, not just headline AI demand. That makes the next 1-3 months about guidance and capex commentary; the 6-18 month question is whether AI translates into durable cash-flow productivity or remains mostly a supplier-led spending cycle.

The contrarian view is that the bubble debate is likely misplaced in the wrong layer. The vulnerable trade is not necessarily the chip leaders; it is the assumption that index-level earnings strength justifies passive ownership of the whole market. If rates stay elevated, the market will eventually discriminate between cash earnings and mark-to-market optics, which should compress the premium on broad indices before it hits the true infrastructure winners.

Key Decisions for Investors

• Long MU on pullbacks into the next print, 3-6 month horizon. Best operating leverage to AI memory tightness; exit if gross-margin guidance or HBM pricing inflects down.
• Pair trade: long NVDA / short INTC for 1-3 months. Expresses winner-take-most in AI compute architecture versus a legacy laggard with weaker direct leverage to the cycle.
• Use SPY or QQQ as the hedge leg against any AI basket long. The index looks more vulnerable than the leaders because reported EPS is being inflated by concentration and non-recurring gains; a small hedge protects against multiple compression if rates back up.
• Watch hyperscaler capex guidance and memory ASPs as the falsifiers. If either rolls over, reduce exposure to the whole AI hardware complex quickly.