This Week in Markets: The AI Cash-Flow Test and a Rotation Into Everything Else, October 30, 2026

The week in one frame

Data as of October 2026.

The S&P 500 fell 0.57% over the five sessions to 7,580.07 for the week ended October 30, and the Nasdaq Composite dropped 2.00% to 25,437.00, according to Morningstar's weekly market update. Underneath those headline numbers, most stocks went up. Of the 878 US-listed companies Morningstar tracks, 542, or 62%, finished the week higher.

An index falling while the median stock rises is a narrow market running in reverse. The losses were concentrated in the largest technology names and the gains were spread almost everywhere else. Morningstar's US Total Market index, which carries far less weight in a handful of companies than the S&P 500, fell 0.42%.

Friday delivered most of the damage. The S&P 500 lost 1.82% on the session to 7,580.07, the Dow Jones Industrial Average 1.09% to 51,770.31, and the Nasdaq 2.86%, its steepest single-day decline since June 5. The Russell 2000 fell 1.66%. Trading was one-directional: 2,230 stocks declined against 947 advancers, a ratio of 2.4 to 1.

The artificial-intelligence trade met its cash-flow test

Three earnings reports in five days did more to reset technology valuations than any commentary this year, and all three failed on the same measure, which was cash.

Alphabet reported October 29 with revenue below analyst estimates, and the discussion that followed centered on whether AI answer tools are beginning to erode the economics of traditional search. The stock posted its worst week since 2011, according to week-in-review coverage. Communication services, the sector that carries Alphabet, fell 5.47% over the week, the weakest of Morningstar's eleven sectors.

Amazon reported the same day with free cash flow negative at 42.4 billion dollars for the third quarter, capital spending held near 220 billion dollars for the full year, Amazon Web Services revenue growth of 39% and earnings per share of 1.89 dollars. The cloud growth was the fastest in five years. It was not enough. The shares closed Friday at 240.35 dollars, down 1.41% on the session.

Intel was the most violent case. Fourth-quarter guidance of 15 billion to 16 billion dollars landed well below consensus, and the stock fell 28.58% to 65.76 dollars, the largest single-day decline in its history, on 450.9 million shares, making it the most heavily traded name on the tape.

The common thread is not that spending rose. It is that spending rose while the revenue it is meant to produce became harder to see. Capital outlays hit the income statement through depreciation before they show up as growth, which is the mechanism behind why extreme valuations compress when the payback period lengthens.

Breadth paid, and the payment went to health care

Health care rose 3.06% and utilities 2.00%, the two best sectors of the week. Technology fell 1.80% for the week and 3.93% on Friday alone. The rotation was visible in style and size as well: large-cap stocks fell 0.63%, mid-caps rose 0.84% and small caps rose 1.08%, while value gained 1.25% and growth lost 1.78%.

Health care climbed for reasons specific to it. It is the least AI-exposed of the large sectors, so it functions as a parking space when the AI trade is being repriced. Utilities are more interesting, because they normally fall when bond yields rise and AI data-center demand has made them a technology-adjacent trade. This week they rose while the 10-year Treasury yield fell to 4.94% from 5.00%, so the bond-proxy effect and the electricity-demand story pulled in the same direction at once. The economics of that electricity demand are covered in the energy strategy behind data centers.

The size result deserves its own line. Small caps outperformed large caps by 1.71 percentage points in a week when the policy rate stayed at 3.75% to 4.00% and inflation ran at 3.6%. That is not what a rate-sensitivity model would predict, and it is the clearest evidence that this week's selling was about specific cash-flow expectations rather than about the cost of capital in general. A cap-weighted benchmark and an equal-weighted one told different stories this week, which is the practical argument in the case for owning the whole index rather than a basket of favourites.

The single-stock tape split into two baskets

The week's ten worst large-cap performers and its ten best had almost nothing in common except that the losers were priced for AI growth and the winners were not.

Reddit fell 36.41% for the week, including 29.22% on Friday to 91.73 dollars, after fourth-quarter guidance came in soft and management flagged a decline in search referral traffic along with a plan to test requiring logins to view the site. ServiceNow lost 33.83%. In Brazil, Banco Bradesco fell 37.35% and Itau Unibanco 26.27% after the presidential election result and a contested concession.

On the other side, Herc Holdings gained 29.22% to 165.54 dollars, Hayward Holdings 27.23%, Wingstop 25.36%, Range Resources 20.01% and Pentair 19.82%. Equipment rental, pool equipment, restaurants, natural gas and water treatment: industrial and consumer businesses with no data-center narrative attached. Expensive growth falling while cheap cash-generative businesses rise is the rotation value investing is built to capture.

One employment signal belongs in this section because it is a cost decision rather than a demand story. Microsoft offered buyouts to as many as 60,000 workers against headcount of 223,000 at the end of June, with the figure nearly flat since 2022, when its data-center spending accelerated. Rising capital budgets have to be funded from somewhere, and operating expense is one of the places.

Rates fell even as hike bets built

The Federal Reserve held its benchmark rate at 3.75% to 4.00% on October 28 by a 10-2 vote, with Governor Lisa Cook and Cleveland Fed President Beth Hammack dissenting in favor of a quarter-point increase. The next day the September personal consumption expenditures index showed a 0.5% monthly rise, the largest in six months, and an annual rate of 3.6% against 3.4% the month before, with core at 3.1% from 3.0%.

The bond market's response was the week's most useful puzzle. The two-year Treasury yield fell to 4.56% from 4.69%, a 13 basis point decline, and the 10-year fell 6 basis points to 4.94%, in the same week that futures pricing moved toward a higher probability of an increase in December. Front-end yields and rate expectations normally travel together. When they separate, one of the two is looking at something the other is not, and this week the bond market was pricing slower growth ahead while the inflation data was arguing the opposite.

That matters for equities through the discount rate rather than through earnings. A lower 10-year yield mechanically raises the present value of distant cash flows, which is why the deepest-duration growth names would have been the last to fall under a rates-led explanation. They were the first, which tells you valuation was doing the work and rates were not. Volatility behaves this way when positioning, not fundamentals, is the binding constraint.

Outside the United States, two decisions added to the same theme. The European Central Bank held its deposit rate at 2.5% on October 29, and President Christine Lagarde said she will step down on February 7, with markets pricing a 67% chance of a December increase. The Bank of Japan held at 1.25% on October 30, revised its inflation forecast higher, more than doubled its annual ETF sales pace to 860 billion yen, and Governor Kazuo Ueda signaled a December move is possible.

Asia's chip complex and Europe's best month since June

The Nikkei 225 closed October 30 at 63,290.11, down 1,369.36 points or 2.12%, led by semiconductor shares. South Korea's KOSPI fell 3.48% on the session, with SK Hynix down 7.08%, roughly twice the index move. The Shanghai Composite closed at 3,599.02, down 1.44%, extending a losing run to a twelfth consecutive week and leaving mainland blue chips with their steepest monthly loss since early 2016.

Hong Kong was the exception in the region, with the Hang Seng up 1.50% on Friday, the widest one-day gap between the two Chinese markets in some time.

Europe had the mirror image of Asia's week. The STOXX 600 fell on Friday as chip names sold off, but closed out its best month since June, and the FTSE 100 finished 7.71 points higher, up 0.1%, at 10,729.78. An index weighted toward banks, miners and energy companies is structurally insulated from a semiconductor repricing, and this week it was. Index construction is not a neutral choice, which the mechanics of passive ownership make clear.

Volatility, gold and what to watch

The Cboe Volatility Index ended at 19.42, up 12.91% on Friday, still well short of the levels that accompany a genuine regime change. Gold fell 2.88% over the week to 3,879.30 dollars an ounce and 3.69% on Friday, which is unusual company for a risk-off week and points to forced selling rather than a shift in demand for safety. West Texas Intermediate slipped 0.41% to 103.02 dollars while Brent rose 1.32% to 112.26 dollars on Friday, a spread of 9.24 dollars between the two benchmarks.

Two data points set the backdrop. The Commerce Department's initial estimate put third-quarter gross domestic product growth at 3.4% annualized, up from 2.2% in the second quarter, with business investment the main driver. The Conference Board's consumer confidence index rose 7.5 points to 91.7, the highest since June, though the survey also recorded concern about prices outpacing wages.

The week ahead has three tests. Apple and Palantir report November 2, which is the first megacap result after the AI repricing and the first read on whether the cash-flow objection extends to hardware and software. The Bureau of Labor Statistics publishes the October Employment Situation report on November 6. And December decisions from the Fed, the ECB and the Bank of Japan now all carry live increase risk, which means the year's final month is priced for tightening rather than for a pause.

The confirmation level is breadth. If the index falls again while advancers keep outnumbering decliners, the rotation is doing its job and the damage stays concentrated in the companies whose valuations depend on AI payback. If advancers and decliners turn with the index, the drawdown has broadened into something that will show up in every portfolio regardless of sector, which is the difference between a repricing and a market that has changed direction. The concentration that made the index look calm before this week is the subject of the American valuation problem.