By Liquiditrax ResearchPublished
The earnings scoreboard looks great on the surface and mean underneath. Both things are true, and the gap between them is the signal.
The breadth is genuinely strong. Of 158 S&P 500 firms that reported, 132 beat EPS (84%), 18 missed, 8 matched, and 113 (72%) grew profit year over year. On revenue, 123 beat and 138 posted YoY sales growth. By the numbers this is a healthy earnings season.
But beating the number stopped being enough. Look at the mega-cap reactions. Microsoft rose 15.5% on Q2 revenue of $90.01B (+17.7% YoY) and Azure +43%. Amazon rose 15.3% on $200.61B (+19.6% YoY) with AWS +36.7%. Then Apple fell 7.4% even though revenue grew 16% YoY, because Services and Greater China missed. Meta fell more than 7% despite 28% revenue growth, because EPS missed and it raised full-year capex guidance to $130B-$145B for AI infrastructure.
The dividing line is capex ROIC, not the headline beat. The market rewarded the names where AI spend is visibly converting to cloud revenue and punished the names where spend is rising faster than the payoff is proven. Exxon slipped 1% even with earnings up 67% sequentially to $14.7B; Ford rose 2% on a smaller beat after raising EBIT guidance. Reaction is about the forward, not the quarter.
Why it matters (allocator lens): a broad beat rate tells you the economy's earnings engine still works. It does not tell you the index goes up, because leadership is being repriced on how efficiently the AI buildout turns into cash. Dispersion this wide inside the mega-caps means the index level hides more than it shows.
So-what for ABC: owning Beta through the index still gives you the healthy breadth, but the return is increasingly a function of which few names carry it. This is the environment where Cash as dry powder matters, because capex-heavy disappointments can gap down hard even on a revenue beat. Don't confuse a strong scoreboard with a green light to add indiscriminately.
Takeaway: 84% beat, and the market still sold the capex-heavy misses. The tape is grading return on AI spend, not the EPS line.
Analytics & education, not advice. DYOR.