#每周来晒 #U.S. Earnings Season

The discussion raised three questions. The first two (which company is most promising, and whether AI investment can continue to grow) have already been debated extensively. But the third question—“Will a stock necessarily rise if earnings beat expectations?”—got an almost textbook answer this earnings season: not necessarily. And the reasons why matter much more than the conclusion itself.

Meta and Tesla offer two cases with opposite outcomes but the same underlying logic. Meta: “beat expectations, but fell.” Revenue, EPS, and Q4 guidance all beat expectations, but the stock fell because FCF turned negative, the capex ceiling didn’t come down, and CDS spreads widened. The market is repricing “the pace of spending.” Tesla: “beat expectations, but also fell.” EPS beat expectations, margins recovered, and FCF unexpectedly turned positive, but the stock fell because capex was cut. The market is repricing “the slowdown in growth implied by not spending.”

Put these two cases together, and the takeaway isn’t the crude conclusion that “beats don’t matter,” but a more precise framework: the market’s valuation of tech giants has shifted from “EPS beat” to “capex narrative + cash flow trajectory.” An EPS beat is merely the price of admission. Two things actually determine the stock’s direction: first, whether your spending (capex) is accelerating or slowing, and whether that direction aligns with how the market values your sector; second, whether your spending is leaving behind an unsustainable cash-flow structure (such as Meta’s negative FCF and financing needs).

This framework also applies to Microsoft, Alphabet, Amazon, and Apple next week, as well as Nvidia in November. If a company beats EPS expectations but its capex guidance falls short of the market’s implicit expectations for an “AI arms race,” it could follow Tesla’s logic for a sell-off. If capex continues to expand but the FCF shortfall grows at the same time, with no clear path to monetization, it could follow Meta’s logic for a sell-off. “Beating expectations” is no longer a tradable signal on its own; what’s tradable is the structure of the beat.

As for “which company is most promising”—under this framework, the answer isn’t any one company, but the one that can minimize the lag between “capex expansion” and “revenue realization.” So far, no company has delivered a complete answer. The real climax of earnings season isn’t in the two reports already released, but in the reports coming next week.