Microsoft +15.5%, Meta −9%: The Week the Market Started Grading AI Spending

The Sunday Setup · Week ending July 31, 2026 · ~6 min read

Last week we asked one question: when Microsoft, Meta, Apple and Amazon opened their books, would the AI rotation keep holding — or would a second wave of capex receipts spread the selling? This week we got the answer, and it was cleaner than anyone expected. The market didn’t reject AI spending. It graded it.

Four of the biggest companies on earth reported within 48 hours. Every single one grew revenue double digits. And the tape sent them in opposite directions. Microsoft ripped +15.5% — the largest single-day gain in market-cap history — while Meta fell 9% and Apple dropped 8%, all in the same window. Same theme. Same spending. Opposite verdicts.

Here’s the shift, and it’s the whole story in one line: the market stopped paying for AI spending and started paying for AI returns. For two years, announcing a bigger capex number was enough to move a stock up. This week that stopped working. The companies that could show the return got rewarded like it was 2023. The ones that could only show the bill got sold — even on a revenue beat. Here’s the promise for the next six minutes: we’ll break the week down the way we’d break down a single setup, and by the end you’ll know exactly why one AI spender gained a trillion dollars while another lost 9% in the same 48 hours.

📊 The Scoreboard

Green across the board — the first up week in three for the Nasdaq. But the headline hides the real signal: this wasn’t a rally, it was a sorting.

★ Green for the week — but the Nasdaq’s +1.6% is the net of two 15% winners and two ~8% losers. And for the month, the Nasdaq still finished −3.2%. July was a grind.

The headline is green, but the story is dispersion. The four mega-caps that reported moved a combined ~$2 trillion in market value in one week — and the direction each went depended almost entirely on one thing: whether investors could see a return on the AI spend. Alphabet, Amazon and Microsoft alone added nearly $1.5 trillion in combined value. Meta and Apple gave a big chunk back.

Read the split. That’s a healthier tape than a week where everything green-candles together on a headline. Green-together is a mood. Sorted is a market doing its job. It took the last two sessions — Amazon and Microsoft — to drag the tape green for the week and let the Dow post its fourth straight monthly gain.

🔍 Decoded: The Two That Actually Matter

1. The market repriced AI from a story to a spreadsheet

Every mega-cap this week grew revenue double digits. Microsoft did $90B (beat), Amazon beat, Apple beat, Meta grew revenue 28% year over year. On the old rules, all four should have been up. Instead the tape split them by a single question: where’s the return?

The lesson hiding in the numbers: revenue growth is now the ante, not the edge. Everyone has it. What the market paid up for this week was proof the AI money is coming back as profit — and it punished, hard, anyone who spent more without showing it.

2. A hawkish Fed hold nobody’s talking about

Buried under the earnings fireworks, the FOMC held rates at 3.5%–3.75% on Wednesday — but the vote was the story. Three members dissented, and they didn’t want a cut. They wanted a hike. With oil pushing higher on Middle East tension, the inflation conversation is quietly turning the wrong way.

📈 Chart of the Week: MSFT Through the Engine

Microsoft reported Tuesday after the close and gapped up roughly 15% the next morning — its largest single-session move ever — clearing the low-$400s ceiling it had been stuck under for weeks, and then held the breakout the following day. A textbook confirmed breakout, and the cleanest “wait for it” trade of the year.

Here’s the sequence, run through the Alignment Engine:

1. Bias — a coiled leader, not yet proven. Going in, MSFT was constructive but capped: strong business, but stuck under a shelf in the low $400s while the whole group waited on the capex verdict. Bias was up, but unconfirmed. “Looks ready” is not “is ready.”

2. Key Level — the low-$400s ceiling. One line framed the entire trade: the resistance shelf around $400–410 that had rejected the stock repeatedly. Below it, dead money. A daily close above it, on a catalyst, changes the character completely.

3. Reaction — the gap that meant it. Wednesday’s open gapped MSFT ~15% straight through the shelf on record volume, with a real fundamental behind it: Azure accelerating to 43%, crossing $100B, guiding higher. A level that breaks on a gap, on volume, with earnings confirming it, is a level that means it.

4. Confirmation — it held and extended. The next session MSFT didn’t fade the gap — it added another +3% to close near $465. The broken ceiling became the floor. That hold is the confirmation the Engine waits for: the market re-rating the name and keeping it there.

5. Execution — buy the proven break, not the pre-earnings guess. The qualified trade was never to gamble a long into the print. It was to let MSFT prove the reclaim above $410 and hold it, then enter with risk defined under the breakout. You gave up the first candle of the gap. You kept your capital off a coin flip and still caught a re-rated leader.

💬 What the Street Is Buzzing About

The timelines split into camps before the reports even landed — extremely bullish on Amazon and Alphabet, cautious on Microsoft, outright bearish on Meta and Apple. Then the numbers hit and the mood whipped around in hours.

🔥 “Microsoft just ended the AI-bubble debate.” When MSFT ripped 15%, retail sentiment flipped to extremely bullish and message volume jumped over 360% in a day. The enthusiasm is earned — Azure accelerating is real proof the spend is converting. But a 15% gap is not an entry. The buzz peaked after the move, which is exactly when the easy money is already gone. We liked MSFT at the confirmed reclaim, not at the top of the timeline’s excitement.

🧊 “Meta’s a falling knife — capex with no payoff.” Bears piled on after the 9% drop and the free-cash-flow collapse. The concern is legitimate; the market is right to demand a return. But “falling knife” is a description, not a trade. Meta found a floor near $521–525 by Friday. The trade isn’t shorting a stock that’s already down 9% or catching it because it “looks cheap” — it’s waiting for it to prove a level, in either direction.

🤝 Inside the Community

Every morning before the bell we run a live pre-market session — walk the indices, mark the levels, build the watchlist, then trade it live. In the biggest earnings week of the season, the room’s edge was what it didn’t do: nobody in the log gambled a single earnings print. With Microsoft, Meta, Apple and Amazon all coin-flipping after hours, the team traded the one thing that stayed readable — the index levels — and let the mega-cap lottery play out without them.

The teaching moment was the same setup, over and over: breakdown-and-retest on QQQ and SPY. While CNBC ran wall-to-wall earnings coverage, the indices kept handing the room clean, mechanical levels — break a shelf, retest it from below, short the rejection. On Wednesday, after the FOMC air-pocket, the retest short on the indices was textbook. The lesson for the room: you don’t need the exciting trade. You need the readable one.

Six tickets, five green and one red — and the red one is the one worth studying. Every entry at a level, every risk defined before the trade, every result logged after.

👤 Member spotlight — the recovery, not the win. Prashanthi traded independently all week — no signals, no hand-holding — running the same disciplined breakdown-and-retest process across three sessions. But the trade worth framing isn’t one of her winners. It’s the loss. Friday morning, her first SPY short stopped out for −1R. Instead of revenge-trading or freezing, she went back to her plan, re-read the level, and took the next qualified setup — booking +2.3R, then scaling another out to +3.4R. She logged the loss honestly, with the fix attached: “stop was too tight — should’ve placed it above PMH, not just above the retest wick.” That’s the whole game. She didn’t let a red ticket become a red day. Green teaches you the setup. Red teaches you the discipline — and how you respond to red teaches you everything else.

📅 The Week Ahead

Data — it’s a jobs week. The calendar flips from earnings to the labor market, and it builds all week to one number. ISM Manufacturing Monday, JOLTS Tuesday, ADP + ISM Services Wednesday — then the main event: the July jobs report Friday morning. After a hawkish Fed with members itching to hike, a hot payrolls number would put rate cuts even further out of reach. This print sets the tone for August.

Earnings: the mega-cap wave is past, but the second tier reports into a market that has clearly shown it will pay for returns and punish spending without proof. The bar just changed. Watch how names guide, not just what they earned.

Oil and the Fed’s headache: WTI is pushing ~$85 and Brent near $90 on Strait of Hormuz tension. Rising energy is the one thing that keeps the hawks loud and the cuts on ice.

The watchlist — levels we’re marking

The real question: does the market keep sorting — rewarding proven returns, punishing unproven spend — or does a hot jobs number turn a selective tape into a risk-off one? Selective is healthy. Risk-off is not. Define your invalidation before the open, and let the reactions come to you. In a jobs week, the move you want is after the number, not before it.

Meta Trading Club provides educational content only. Nothing here is financial, investment, or trading advice, or a recommendation to buy or sell any security. Trading involves substantial risk of loss. Past performance does not guarantee future results.