Investing / The Compound Effect
Microsoft Gained $450 Billion in a Day. Apple Lost $475 Billion the Next. Same Story.
| 4 min | By Heath J. Harris
The AI arms race stopped being a spending contest this week and became a grading period. Four earnings reports, one hawkish Fed, and the sorting Wall Street has been putting off all year.
This was the week the market stopped paying for AI promises and started grading AI report cards. Microsoft added roughly $450 billion of market value on Thursday, the largest single-day gain by any stock in history (Yahoo Finance, July 30, 2026). Apple erased about $475 billion on Friday (CNBC, July 31, 2026). Two record-scale moves, twelve hours apart, driven by the same force pointed in opposite directions.
If you own an index fund, you own both sides of that trade, which is why this week deserves ten minutes of your attention.
First, the Fed picked a family fight
The Federal Reserve held its target range at 3.50% to 3.75% on Wednesday, the fifth straight meeting without a move. The vote was 9 to 3, and here is the part that matters: all three dissenters wanted a hike, not a cut (CNBC, July 29, 2026). Chair Kevin Warsh, whom we flagged in this newsletter last week as the man to watch, summed it up afterward: "I asked for a good family fight, and I got one" (Forbes, July 29, 2026).
The bond market took the hawks' side. The 10-year Treasury yield touched 4.74% on Friday, its highest level since January 2025, and the 30-year pushed above 5.2%, territory it has not seen since 2007 (Trading Economics; TheStreet, July 31, 2026). Stocks noticed. The Dow fell 1,153 points on Wednesday, a 2.2% drop and its worst session since April 2025 (CNBC, July 29, 2026).
The data explains the fight. Headline PCE inflation came in at 3.7% for June, down from 4.1% but nowhere near the 2% target, while the advance estimate put second-quarter GDP growth at just 1.5% against expectations of 2.1% (BEA, July 30, 2026). Cooling growth and hot prices at the same time. There is no textbook chapter the Fed likes less. One more number worth filing away: inflation-adjusted private-sector wages fell 0.4% over the past year, the first real-wage decline since 2022 (BLS, July 31, 2026).
Then the report cards came out
Microsoft earned its record day. Revenue rose 18% to $90 billion, Azure grew 43%, and Azure crossed $100 billion in annual revenue for the first time (CNBC; Microsoft, July 29, 2026). The AI spending shows up on the bill too, at roughly $41 billion of quarterly capital expense, up 69% from a year ago (Axios, July 29, 2026). The market looked at the spending, looked at the revenue growing beside it, and paid up 16%.
Amazon told the same story. AWS grew 37%, its fastest clip in 18 quarters, and the stock jumped almost 15% on Friday even as the company raised its capital spending plan to roughly $220 billion (CNBC; Motley Fool, July 31, 2026).
Meta spent the money without showing the same receipts. Revenue beat, but earnings missed, free cash flow shrank to $784 million for the quarter, and the capex budget grew again to as much as $145 billion. The stock fell about 10% (CNBC, July 29, 2026). Apple grew revenue 16% to its strongest June quarter ever, then warned that the AI buildout is squeezing the memory chips and components it buys, raising its costs into the fall. Down 9.4% (CNBC; 24/7 Wall St, July 31, 2026).
Add it up and Big Tech's combined AI budgets now run past $600 billion a year (Axios, July 29, 2026). For two years the market funded that arms race on faith. This week it started separating the companies that turn AI spending into revenue from the companies that just spend, and from the companies stuck buying components in the middle of it.
What the sorting means for your money
Watch what the whole week did, not any single day. The S&P 500 finished with its first winning week in three (Motley Fool, July 31, 2026), and July as a whole closed slightly negative. Trillions of dollars of single-stock drama netted out to a market that ended the week higher and grumpier.
That is what sorting looks like, and we expect more of it. When the index giants move 10% to 16% in a day on earnings, a cap-weighted fund quietly reshuffles its own top ten, and a portfolio that was at its target weight in June may not be at its target weight today. The useful response is not a prediction about which stock wins the AI race. It is knowing exactly how much of the race you own, and whether that number still matches your plan.
The sorting also runs on rails the Fed controls. With three dissenters pushing for hikes and long yields at multi-year highs, the price of patience just went up for every company spending before the payoff. Hold that thought until Nvidia reports in late August.
As always, this is education, not advice.