←2026-07-29 · VM Genius
Analysis
Last night, the Dow surged 500 points, posting a respectable 1% gain; at the same time, the Philadelphia Semiconductor Index cratered 4.5%, and Meta suffered its longest nine-day losing streak since its IPO in 2012.
On the eve of “Judgment Day”—when both the Fed decision and the Microsoft and Meta earnings are set to land—where did the smart money go?
Numbers don’t lie.
The S&P 500 closed up 0.2%—looks like another boring day, right? Wrong. The S&P 500 rose just 0.2%, but the S&P 500 equal-weighted index surged 1.1%. The “common stocks” within the index were rallying in unison, and the only thing dragging the index underwater were those few AI behemoths with the largest market caps.
The Dow gained 1% (up over 500 points), the Russell 2000 added 0.2%, and healthcare, consumer staples, and materials all closed higher; on the other side, the Nasdaq 100 fell 1%, now sitting about 9.5% below its late-June all-time high—one foot already on the doorstep of a “technical correction.” The Philadelphia Semiconductor Index tumbled 4.5%, its fourth straight losing session—a losing streak last seen back in March. Meta fell for a ninth straight day, the first time in its fourteen years as a public company.
The most unusual detail: Treasury yields have fallen for three straight trading sessions, with the 10-year back to 4.60%. According to the old playbook, when rates fall, tech rallies. But this time, rate-cut expectations couldn’t buy the Nasdaq 100 even a single day of relief. When falling rates can’t save tech stocks, it means the market isn’t selling the “discount rate”—it’s selling the “story itself.”
Microsoft, Meta, Amazon — behind each of them lies a cash cow. Over the past four or five years, they’ve used the cow’s milk to irrigate the largest infrastructure buildout in human history: an unprecedented mega-factory. The problem is, nobody knows what this factory is producing, and even less do you know who’s going to pick up the tab.
When financial data starts to prove the input-output ratio is flunking, who will be the first to pull back from the brink? Meanwhile, Apple, which has been watching from the sidelines, could end up the biggest winner. The players who jumped in too early and splashed out too much cash might have to hand over 10% to 20% of their market cap as tuition.
And then there are the smaller companies — the ones without a cash cow, financing data centers purely with debt. Once they fall, they’ll drag down private credit, pulling everyone else under with them. That’s the full puzzle of the “circular financing” fear: the giants’ capex is the little guys’ revenue, the little guys’ debt is the asset of the credit market. Everything is linked; one loose link, and the whole chain gets repriced.
Kristina Hooper, Chief Market Strategist at Man Group, delivers another blow: investors have already executed a 180-degree turn. Last year the market rewarded companies that dared to spend; this year it’s punishing them — especially those borrowing to spend. Hyperscale cloud providers could face a wave of repricing, and the same goes for semiconductors.
Take note: it’s a hike, not a cut. This may be the most underestimated risk this week.
Tomorrow at 2 p.m., the Fed announces its decision — the second meeting since Kevin Warsh took office. The probability of a hike on Wednesday is around 30%; for September, that figure has surged to 60%–70%.
Former Fed Governor Betsy Duke’s comments carry the most weight: she doesn’t expect a change this time, but “the case for a rate hike is very strong.” The inflation target hasn’t been met for more than five years, the labor market is stable, growth is solid, and rates aren’t even restrictive. Having thrown down tough talk about controlling inflation, doing nothing now would be a genuine credibility crisis.
There’s also a view that Wednesday will most likely be a “hawkish hold,” with the real window for a hike opening in September. Why? When Warsh was nominated in January, the San Francisco Fed’s “proxy federal funds rate” had already decoupled to the upside from the actual rate — in plain terms, the market had long priced in “Warsh will shrink the balance sheet and tighten.” That gives him enough political capital to stand pat and watch for now.
For tech stocks, this is the second sword hanging overhead. Tech stocks are classic long-duration assets — any rise in long-end rates or an actual hike presses directly on valuations. On one side, the earnings trial; on the other, the rate trial. Both swords fall together on Wednesday.
Here’s the irony: on the very same day AI faith got stress-tested, the “old economy” delivered one blowout report after another—
Coca-Cola surged 5%: raised its full-year outlook, posted 5% organic sales growth, lifted its EPS growth guidance to as high as 8%, and saw North America organic sales jump 7% — money spent on World Cup sponsorship clearly paid off. The stock is up 26% year-to-date.
Boeing jumped 4.8%: second-quarter free cash flow came in at $631 million, whereas analysts were bracing for an outflow of $331 million — a swing of nearly $1 billion.
Ford gained 4% after hours: adjusted EPS of $0.42 crushed the $0.36 consensus, and full-year EBIT guidance was raised to $10 billion–$11 billion. The playbook is brutally simple: sell fewer money-losing EVs (EV sales tumbled 41%) and sell more high-margin pickups and SUVs priced between $60,000 and $80,000. The upper half of the K-shaped economy is still happy to buy big trucks.
Seagate rallied 7% after hours: next-quarter revenue guidance of $4.0 billion–$4.2 billion blew past the Street’s $3.79 billion estimate. The hard-drive “old business” is actually riding the AI data deluge.
Mondelez rose 1.5%: both revenue and profit beat expectations — Oreos remain hard currency.
Of course, not everyone aced it: Caesars missed big on adjusted EBITDA. Visa beat across the board — EPS, revenue, and transaction counts — yet the stock slipped over 1% after hours, because the market zoomed in on something else entirely: the company that never lays people off announced it is cutting 7% of its workforce and booking a $563 million charge.
Visa states that consumer and business spending remains resilient, with payment transaction volume growth accelerating from 9% to 10%. UPS CEO Carol Tomé provides the most vivid footnote: UPS handles roughly 3% of U.S. GDP in freight volume. "Where the consumer goes, the economy goes," and the consumer she sees remains tenacious. By the way, UPS's most profitable international lane — China to the U.S. — surged 22% in Q2, and intra-Asia lanes grew 14%. The capillaries of global trade are regaining blood flow.
Yet UPS shares fell 6% that day. The earnings clearly beat expectations, and guidance was raised — so why the drop? The market is nitpicking the second-half revenue growth deceleration from 9% to 4.5%. Tomé explained it away as a base effect from Amazon voluntarily pulling back volume; strip that out and growth continues. But the market is merciless — it would rather kill first than wait. Beating expectations is no longer enough; you have to beat them perfectly, in every dimension, without a single blind spot.
If you can only remember one detail from last night, make it this one.
Insurers—the canary in the coal mine—have already started excluding AI risk from corporate policies, or doubling their premiums.
Why does this signal matter? Insurers are the first in the market to price risk. In 2020 and 2021, cybersecurity was carved out of general liability policies and priced separately. Now the same script is playing out with AI, and this time it’s worse: cybersecurity guards against “outsiders coming in,” but AI risk guards against “insider data going out.” For the first time ever, companies are actively encouraging employees to use frontier models outside the firewall—effectively opening their own back door.
While the capital market is still debating AI’s ROI, insurance actuaries have already voted with their premiums. The lesson from history: the canary knows the answer before the miner does.
The first light: it’s rotation, not collapse. Last night wasn’t a 2022-style indiscriminate rout: the equal-weighted S&P 500 surged 1.1%, most stocks were up, money didn’t leave the market—it just switched its faith. This is the internal metabolism of a bull market, not its funeral.
The second light: a math problem. Bonds are very cheap relative to stocks. Either equity prices come down, or bond prices go up—either way, anyone holding long-duration bonds doesn’t lose. At the margin, capital is pulling out of the most expensive stocks and buying cheap duration.
The third light: “collective overinvestment,” not a death sentence for all. Someone will always step back from the cliff’s edge, someone will always end up as an Apple-style winner sitting on the sidelines. Bubble theorists and believers agree on one thing: AI is a genuine sea change—what they’re fighting over is just the price of admission.
The market, with four consecutive days of chip selloffs and Meta’s nine-day losing streak, is telling us: the era of free-lunch valuations is over. AI companies now have to let ROI do the talking.
Money is moving out of the most expensive beliefs and into Coca-Cola, airplanes, pickup trucks, and bonds. The equal-weight index’s 1.1% gain is the receipt for that relocation.
Tonight (Wednesday, Eastern Time) is the real judgment day: the Fed decision at 2 p.m., Warsh on stage at 2:30 p.m., and after the close Microsoft, Meta, and Qualcomm all dropping earnings at the same time. Either the hyperscalers prove with numbers that the factory can mint gold, or the market continues to vote with its feet. One sword is rates, the other is earnings — only a night when both swords fall at once deserves to be called judgment day.