Apple and Amazon close out the gauntlet: how the crowd read the first two, and what it expects from the last two

Yesterday we wrote that nearly every account in the room was reciting the same script going into last night: beat, raise capex, dump anyway. Twelve hours later the script met the numbers, and the verdict split clean down the middle. One name broke it. One fulfilled it with room to spare. Tonight, after the close, Apple and Amazon report and close out the biggest week of the season, and they walk in looking nothing alike.

The script met the numbers

Here is how the two rooms moved between the prints landing last night and this afternoon, in the accounts we track:

Going in, both rooms were quiet and undecided. Microsoft was 50% bullish on the week, Meta was split 41 to 43. There is nothing undecided about either room now.

Microsoft broke the script

Microsoft beat on essentially everything. Revenue of $90.0 billion against an $87.7 billion estimate, Azure up 43% against an expected 40%, and Azure crossing $100 billion in annual revenue for the first time. The first reaction in our set, two minutes after the print: "$MSFT +4% initial reaction." By this afternoon one account was posting "$MSFT is up 14% after crushing earnings," double the 7% move the options market was pricing going in. The stock did not just beat and hold. It beat and ran.

The room's more careful readers still found the caveats. One walked through the fine print: earnings per share got roughly $0.30 of help from a one-time $3.2 billion gain on the company's Anthropic investment, capex hit $41 billion for the quarter, up 70% year over year, and free cash flow fell 23% paying for it. The spending half of the script was real. The dump half never came, because the AI spending showed up on the revenue line too: one widely shared read put it as "the question is no longer whether AI demand is real. Microsoft just showed it's becoming AI revenue."

That reading matters beyond Microsoft, because it is the demand signal for the AI-infrastructure names where, as we wrote yesterday, the trusted rooms have been loud and one-sided all month. Those rooms took the print personally. One account, who spent the week rotating out of Microsoft into memory names, put it plainly: "Saved semis too." On our board this afternoon the receivers are still where the conviction lives: Nebius at 201 mentions on the week, 72% bullish, score 79. Micron 68% bullish at a score of 85. Nvidia at 134 mentions and a score of 80.

Meta followed the script, and then some

Meta's revenue actually came in fine, $60.8 billion against $60.2 billion expected. But earnings per share landed at $6.18, down 13% from a year ago, weighed by $2.4 billion of legal charges and $1.2 billion of severance from the May job cuts, and the full-year capex range narrowed from $125-145 billion to $130-145 billion. A raised floor is still a raise, though one of the room's calmer voices argued the opposite read: "that's guidance getting more precise, not more aggressive."

The market went with the first read. Shares fell 7.6% in post-market trading, and by lunchtime today the room was talking about Meta in the low $500s. What sealed it was the one question the whole call was waiting on. Before the call, one account wrote: "Dear $META, please say the following magic words in ~45 minutes and watch your stock go much higher. '2027 will mark a peak in CapEx.'" An hour later, the same room, quoting the CFO: "No guidance on 2027 CapEx." The next morning brought a wall of price-target cuts, and the crowd's verdict shows in the table above: 56% bearish since the print, with the score falling from 61 to 38.

Put the two halves together and yesterday's setup resolved with unusual symmetry. The two back-to-back posts we quoted, one insisting "whenever they say capex stock goes down," the other "capex is priced in, big pump incoming," were both right. Each just had the wrong name. That is what priced in actually means: the same headline, spending more on AI, was punished where the revenue link is still a promise and rewarded where it showed up in the numbers.

Tonight: one loud, one nearly invisible

Which brings us to the two names reporting after the close, confirmed after-market in our events data. They could not be walking in more differently:

Apple is having its loudest week in our coverage. On Monday we counted 23 mentions and a score of 81. The room has since tripled to 70 mentions, swelling through the stretch where Apple passed Nvidia to become the most valuable public company and then briefly touched a $5 trillion market cap. But look at what the extra attention is made of: the bullish share went down, not up, and the biggest single day in the chart below is mostly non-directional, milestone chatter and passing references rather than actual calls. The score has cooled from 81 to 45 while the headlines got louder, which is the machine telling you the new crowd arrived to spectate, not to argue a thesis. The room's own skeptics said it in words: "$AAPL added $1T in market cap in just the last 22 trading days. Earnings in a couple days. I gotta imagine name is priced for perfection here..."

One more Apple note. The company has historically run the largest share buyback program in the market, and buyback headlines are exactly the kind of announcement crowds cheer reflexively. If tonight's report brings another one, we published a guide to judging whether a buyback actually helps you today.

Amazon is the opposite story. Twelve direct mentions in seven days, from a room that was at 11 on Monday. It shows up in our set constantly, but almost entirely as scenery in other stocks' threads, the wall of passing references in the chart below, not as the subject of anyone's thesis. With a room this small the 33% bullish, 50% bearish split is weather, not climate, though the score has warmed from 37 to 51 since Monday. The one substantial bull case in the set this week argued the stock has "corrected nearly 10% over the past two weeks" and is worth holding through the print. The room's mood is closer to another post from Tuesday night: "I really need Amazon to at least run up to earnings. I don't have the balls to hold through earnings."

What we will be watching

The options market is pricing a move of about 4.5% for Apple tonight and about 7% for Amazon, per an account tracking implied moves early in the week. Worth remembering that Microsoft's implied was 7% and it moved double that. And the script has one name left in it: the most credible account in yesterday's piece wrote "let me guess, MSFT META and AMZN beat on everything but increase capex and dump." Amazon is the last test of that sentence.

As for the Fed decision that landed between the prints: in the 24 hours since, 34 of the roughly 800 signals in our set mentioned the Fed or rates at all, about one in twenty-four. The room shrugged at the macro event of the month and spent the day arguing about two earnings reports, which tells you what this crowd actually trades. Same drill as last night: the useful signal is not tonight's first reaction but whether the credible voices, the ones with a track record, hold their read once the numbers land. Tomorrow we close the week with the full scorecard.

Quantral surfaces signals and context from public sources to support your own research. Nothing here is financial advice or a recommendation to buy or sell.