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 expects S&P 500 EPS to grow 50.4% in Q2 (fastest since Q2 2021’s 91.6% rebound). The article argues the earnings surge is supported by real AI infrastructure economics—e.g., Micron’s GAAP net income rose from $1.89B to $28.2B (up >10x)—and highlights forward P/E of ~20 as partly offsetting valuation/CAPE concerns. Overall, it frames the AI-driven earnings momentum as likely to sustain market upside, even as bubble/“circular financing” and potential chip price declines are debated.

Analysis

The cleanest read-through is that the AI capex complex is still monetizing faster than skeptics expect, but the profits are highly concentrated. NVDA and MU are the direct beneficiaries; the second-order winners are the hyperscalers, especially AMZN and GOOGL, because equity-method gains and cloud demand can boost reported EPS even before operating leverage fully shows up. That means the headline index multiple can stay supported even if breadth remains poor, but it also means the market is leaning on a narrow set of balance sheets and accounting marks.

The biggest near-term risk is that investors confuse a cyclical earnings pop with a durable secular step-up. Memory and accelerator economics are the first places where supply response can catch up, so MU is the most vulnerable to margin normalization over 1-3 quarters if pricing peaks; NVDA is more insulated, but still exposed if hyperscaler capex growth decelerates into 2026. INTC is the structural loser in relative terms because this cycle rewards specialized compute and HBM-linked supply chains rather than legacy CPU share.

Contrarianly, the consensus is probably underestimating how much of the current EPS surge is index-level noise rather than broad profit growth. If the AI build-out is still early, that helps the enablers; if it is merely peaking in spending intensity, the next phase is multiple compression, not earnings collapse. The key falsifier is a capex-guidance reset from the hyperscalers or memory pricing rolling over before the next earnings season; either would hit the trade within weeks, not years.

Key Decisions for Investors

• Long NVDA / short INTC for 3-6 months: express the view that AI spending is rewarding specialized compute while legacy x86/foundry optionality remains less monetizable. Risk/reward is attractive if NVDA guides through the next print while INTC remains margin-constrained; cut if INTC shows credible AI share gains or NVDA gross margin compresses.
• Overweight MU only tactically into the next earnings window, then reassess: the stock should keep outperforming while HBM pricing and utilization stay tight, but it is the fastest to mean-revert if memory supply normalizes. Use a trailing stop keyed to management guide-down risk rather than price alone.
• Pair long AMZN / GOOGL versus broad market via QQQ or SPY hedge: the market may be underpricing how much AI-related equity gains and cloud demand can support reported EPS, but the beta should be hedged because the broad index is already expensive. Best entry is on any post-earnings pullback, with a 1-3 month horizon.
• Avoid chasing the S&P 500 as a pure index long at current valuation; prefer a barbell of AI infrastructure winners and defensive cash flow names. The risk is not an earnings recession, it is narrower leadership and multiple vulnerability if the AI capex cycle slows.
• Set an alert on hyperscaler capex commentary and memory ASPs: if either rolls over, rotate out of MU first and then trim NVDA. That is the highest-signal falsifier for the early-cycle thesis.