Amazon surged 13% after showing exactly what the market has been demanding for three weeks: AI spending that converts into revenue. AWS cloud growth accelerated. Margins held. Free cash flow was positive. Sitting right next to it on the earnings calendar, Apple fell 7.2% after Services revenue and China both missed — the $5 trillion company that doesn't need AI to grow just showed that its non-AI business is slowing. Korea's Kospi rallied 18% in a single session — the largest daily gain in its history — after SK Hynix hit the 30% daily limit and Samsung's results confirmed chip demand is real. The S&P gained 0.7%. The Nasdaq rose 1%. The Dow added 278 points and posted its fourth straight winning month. The 30-year yield sat at 5.26% — still at levels not seen since 2007. The 10-year touched 4.74%. And July — the month that broke the AI trade, hit $100 oil, watched three Fed officials vote to hike, and saw the Nasdaq nearly correct 10% — ended on a rally. Not because the problems are solved. Because the market just figured out who solves them.
Green close to end a wild month. The S&P gained 0.7% to about 7,490. The Nasdaq rose 1% to about 25,373. The Dow added 278 points to 52,486 — its fourth consecutive monthly gain. Chips rallied hard: the semiconductor ETF gained 3.3% on the back of the Korea surge and Amazon's cloud beat. Oil refiners closed at or near all-time highs — PBF Energy is up 170% this year. The war-economy trade is printing money.
| The Numbers I Circled | At the close, July 31 · Day change |
|
| Amazon | AWS accelerated | +13% |
| Apple | China + Services miss | −7.2% |
| 30-Year Yield | 5.26% | since 2007 |
| S&P 500 Sectors | Day change |
|
| Consumer Disc. | | +2.5% |
| Info. Technology | | +1.5% |
| Industrials | | +0.8% |
| Materials | | +0.6% |
| Energy | | +0.5% |
| Financials | | +0.4% |
| Utilities | | +0.2% |
| Consumer Staples | | +0.1% |
| Health Care | | −0.2% |
| Real Estate | | −0.5% |
| Comm. Services | | −0.8% |
| | Notable Gainers | Day change |
|
| |
| |
| |
| |
| |
Amazon was the headline. Revenue beat across every line. AWS cloud growth accelerated — matching the signal Microsoft sent Thursday. The company showed AI monetization in advertising, logistics, and cloud computing. Free cash flow was positive. The stock surged 13% and gave the market its second confirmation in two days that the hyperscaler spending works — if you're the right hyperscaler.
Apple was the opposite headline. Services revenue missed. Greater China missed. The forward outlook disappointed on component costs and demand. The $5 trillion crown from Tuesday is already wobbling. Apple doesn't play the AI capex game the way Microsoft and Amazon do — its business is devices and subscriptions, not cloud and data centers. Today the market punished it for exactly that: in a world where AI monetization is the test, Apple didn't have an answer. The stock fell 7.2%.
Korea told the global story. The Kospi surged 18% — the biggest one-day rally since the index was created. SK Hynix hit the 30% daily limit in Seoul. The chip correction that sent Korea down 6% to 9% in repeated sessions throughout July reversed in a single morning on Samsung's strong results and the Amazon-Microsoft confirmation that AI demand is accelerating. Europe's Stoxx 600 hit a record. The global market decided the AI trade isn't over — it just changed shape.
The 30-year yield settled near 5.26%. The 10-year touched 4.74% intraday — the highest since January 2025. Warsh said this week the Fed has "no magic wand." The bond market agrees, and it's pricing accordingly.
What The Market Is Pricing In
July was the month the market stopped buying "AI" and started buying "AI that makes money." That sounds obvious. But for six months before this, the market treated every company touching artificial intelligence as the same trade — chips, cloud, software, hardware, memory, equipment. If the word AI appeared in the earnings call, the stock went up. July changed that. The month taught the market to read a cash flow statement.
Microsoft showed positive free cash flow alongside record AI revenue. The stock surged 15%. Amazon showed AWS acceleration and positive free cash flow. The stock jumped 13%. On the other side: Alphabet raised capex and fell 7%. Tesla burned cash and fell 14%. Meta's free cash flow dropped 91% and the stock lost 9%. Intel doubled its earnings estimate and still got sold. The market stopped treating these companies as one basket and started grading each one individually.
Amazon surged 13% and Apple fell 7% on the last day of July, and the month's verdict is clear: the AI trade survived, but it evolved — from a sector bet that rewarded spending into a stock-picker's market that rewards cash flow, and the companies that can show AI revenue on the income statement are pulling away from the ones that can only show AI costs on the balance sheet. Microsoft and Amazon are the winners. Alphabet, Tesla, and Meta are on probation. Apple is in a different category — not an AI spender, not an AI earner, just a company whose core business is slowing. The month that started with the SOX entering a bear market ends with the SOX rallying on the strongest Korea session in history. The chips aren't dead. They're just expensive — and the market now knows which ones are worth the price.
After the dot-com crash in 2001, the survivors emerged. Amazon, eBay, and later Google proved that the internet was real — the revenue models just needed time. The pretenders disappeared. The technology survived. The stocks sorted themselves. July 2026 was the same kind of month for AI. Not a crash. A sorting. The demand is confirmed by every company that reported. The spending is enormous. The question — the only question — is who converts the spending into earnings. This month, two companies answered. Next month, the market will ask everyone else.
Three things I'm watching next week:
01 — ISM Manufacturing PMI Monday August 3
The first read on July factory activity. June was 51.7 — barely expanding. July's number captures the full impact of $90-$100 oil, the blockade, the tariff reset on 60 countries, and the memory chip correction that sent Korea into circuit breakers. If the PMI holds above 50, the economy is growing through the war and the 88% earnings beat rate has legs. If it dips below 50 — contraction — the GDP miss at 1.5% was the warning shot and the stagflation conversation moves from theory to data.
02 — Does the 30-year yield stay above 5.20%?
The bond market just posted its loudest week of the year. The 30-year hit 5.26% — highest since 2007. The 10-year touched 4.74%. The market is telling the Fed it's behind on inflation. If long-term yields hold above 5.20% into next week despite the oil pullback and the Iran pause, the bond market is pricing a structural inflation problem that won't go away with one good CPI print. If they ease below 5.10% — on falling oil and peace-talk progress — the September hike odds drop and equities get room to rally. The 30-year is the number that controls everything right now.
03 — Jobs report Friday August 7
Nonfarm payrolls for July. The labor market has been the economy's backbone — unemployment near 4%, hiring steady, wages growing. If July jobs come in strong with sticky wages, the Fed hawks have the full package: hot inflation, hot labor, $90 oil. The September hike becomes a lock. If jobs soften — if the hiring slows or unemployment ticks up — the GDP miss starts to look like a trend, and the Fed faces the worst hand in the book: a slowing economy with rising prices. That's stagflation. The jobs number tells you which road we're on.
July started with $83 oil, the SOX at all-time highs, and a CPI that said inflation was falling. It ended with $100 oil, a 30-year yield at 5.26%, three Fed officials voting to hike, and a market that stopped buying "AI" and started buying "AI that makes money." The trade didn't die. It grew up. Next week the data tells you whether the economy grew up with it.
That's it for today. Have a good weekend. I'll be back on Monday after the close.
— Tom Hartley
Today In Perspective · Published daily, Monday–Friday, after the close
This newsletter is for informational purposes only and does not constitute investment advice. The author is not a registered investment advisor. Past performance does not guarantee future results. Consult a qualified financial professional before making investment decisions.
