Friday, October 30, 2026, written at 9:55 a.m. Eastern, minutes after the opening bell on the last trading day of the month. Nasdaq futures came in down about 2.2% and the S&P contract off 1.1% while Dow futures were flat, which tells you most of the story before you read a single line of it. Amazon is down roughly 7% and Intel down about 25%. The Employment Cost Index landed in line at 8:30 and Chicago PMI came in soft at 9:45. Marks below are from the open and will move.
Top Call: The Demand Question Is Closed, and It Did Not Save Anyone
Wednesday night Microsoft and Alphabet ran the experiment against each other and the market paid the one that spent less. Last night Amazon ran the only version of that experiment that was still missing, the one where a hyperscaler delivers accelerating demand, record operating margin and an enormous backlog all at once, and the answer came back the same. Amazon Web Services grew 39.3% to roughly $46 billion, which is an acceleration, carried a $578 billion backlog, and management guided AWS growth above 40% for both 2027 and 2028. Group operating income hit $28.1 billion at a record 13.9% margin. The stock is down about 7%.
What did the damage is one line. Gross capital spending in the quarter was $77.3 billion against something near $58.7 billion that the Street had modelled, and trailing twelve month free cash flow has now swung to an outflow of $42.4 billion. Management reaffirmed roughly $220 billion of cash capital spending for 2026 and said 2027 will be meaningfully higher. Fourth quarter operating income guidance came in about $2 billion light. Put those together and you have a business whose demand is better than anyone forecast and whose cash position is worse than anyone forecast, and the market now prices the second one.
This is the third consecutive session confirming the same rule, and at this point we should stop calling it a repricing and start calling it a regime. Across Microsoft, Alphabet and Amazon the cloud growth rates were 48%, 103% and 39.3%, every one of them at or above guidance, and the aggregate backlog across the three runs well past a trillion dollars. Nobody disputes the demand any more. What the market is doing is discounting the equity claim that sits behind the spending, and an equity claim behind a build funded out of a widening deficit at a 4.9% ten year is worth less than it was in August no matter how good the order book looks.
The shape of the month makes the point better than any single print. The Nasdaq 100 is heading for its worst monthly underperformance against the equal weight S&P since 2002, with more than $1.2 trillion of index market value erased, while energy is up close to 12% this month and staples, utilities and healthcare are each up more than 9%. The S&P 500 itself is still positive for October. That is not a market deciding artificial intelligence is a fiction. That is a market deciding it will no longer pay a premium multiple for a cash flow profile that looks like a utility build, and reallocating to things that already generate cash today. We would keep leaning that way into November, and we would be very careful about treating any bounce in the mega cap complex as the end of it, because nothing in last night’s results fixes the funding arithmetic.
Amazon: Better Business, Worse Stock
| Metric | Consensus | Actual | Result |
|---|---|---|---|
| Revenue | $202.3bn | $201.9bn, up 12% | Narrow miss |
| EPS | $1.96 | $1.89 | Miss |
| AWS growth | mid 30s% | 39.3% to ~$46bn | Beat, accelerating |
| Gross capex | ~$58.7bn | $77.3bn | Far above |
| Trailing FCF | n/a | minus $42.4bn | The problem |
Two details inside this print matter more than the headline. The first is that inference now accounts for well over half of AWS artificial intelligence revenue, which is the single most bullish structural datapoint of the quarter and almost nobody traded on it. Training revenue is lumpy, contract driven and front loaded. Inference is recurring consumption that scales with how many people actually use the applications, and a business whose AI revenue mix has tipped to inference is a business with a far more durable revenue line than the market is currently crediting. The second is that advertising missed, which is the second company in three days to show softness in the ad line, after Alphabet. One miss is noise. Two in the same week in the two largest digital advertising franchises deserves a second look, and we will be watching Meta’s next datapoint carefully.
Our take is that Amazon is now the clearest example of the gap between a good business and a good stock. We would rather own this franchise than most things in the index on a three year view. We would not step in front of the fourth quarter capital spending disclosure, because the company has told you it is going higher and the market has told you three times this week exactly what it does with that information.
Intel: Where the Shortage Becomes Someone Else’s Problem
| Metric | Consensus | Actual | Result |
|---|---|---|---|
| Q3 revenue | ~$16.1bn | $16.51bn, up 21% | Beat |
| Q3 adj EPS | ~$0.44 | $0.52 | Beat |
| Q4 revenue guide | ~$17bn | $15bn to $16bn | Large miss |
| Q4 adj EPS guide | $0.43 | $0.25 | Large miss |
| Data centre | n/a | Record | Not the issue |
Intel beat the quarter, posted record data centre growth and lost a quarter of its value, which on its own would read as another capital markets story. It is not. The reasons management gave were a personal computer market slowdown that the chief financial officer expects to run at least through the first half of next year, and supply constraints on the data centre components that carry the good margins. That second reason is the one worth sitting with.
For three days we have been writing admiringly about memory being sold out through 2029, about agreed 2027 pricing, about 80% operating margins at SK hynix. Every one of those sentences describes an input cost for somebody. On Thursday Samsung’s own mobile division lost roughly 800 billion won because it has to buy chips from Samsung. Last night Intel told you it cannot get enough of what it needs to build the products that make its money, and simultaneously that the personal computer market is slowing, which is what happens when the bill of materials for a laptop rises faster than anyone is willing to pay. The scarcity rent we have been celebrating at the top of the chain is turning into a margin and volume problem for everyone underneath it, and the market is only now starting to price that second order effect.
This is the piece we underweighted. We had the funding argument right and we flagged the Samsung mobile loss as a warning on Thursday, but we treated it as a curiosity rather than carrying it forward into the names that buy memory rather than sell it. Intel is the first large print where component scarcity, not financing, is the stated cause of a guidance cut. It will not be the last, and the screen worth running this weekend is anyone with heavy memory or substrate content and no pricing power.
Korea Breaks, and It Tells You Who Owns This Trade
The Kospi fell 3.48% to 5,730.74, after being down more than 5% intraday, and the Korea Exchange tripped a sell-side sidecar that suspended program trading for five minutes when KOSPI 200 futures fell 5%. SK hynix closed down 7.08% at 1.42 million won and Samsung Electronics down 3.02% at 225,000 won. Foreign investors sold a net 1.62 trillion won. Both chipmakers are now down more than 16% this month.
Hold that against the fundamentals of the same week. SK hynix reported an 80% operating margin and claimed its books are sold out through 2029. Samsung posted the largest operating profit in its history. Korean industrial production came in much better than expected this morning. Nothing about Korea got worse. What happened is that the Kospi functions as a leveraged claim on American hyperscaler capital spending plans, and when the market reprices the equity that funds those plans, the derivative moves further than the underlying. An index that rose 6.6% on an SK hynix print on Tuesday fell 3.5% on an Amazon print on Friday, and neither move was about Korea.
The practical read is that the memory complex has become a positioning problem rather than a fundamental one. The earnings are real, the sold out order books are real, and the shareholder register is full of people who bought the narrative in the last six weeks and are now being asked whether they want to own a cyclical at the top of a capital spending cycle whose financing has just been questioned three times in a week. We would expect that sorting to take longer than one session.
Macro: A Rate Rally You Should Not Enjoy
| Release | Expected | Actual | Read |
|---|---|---|---|
| Employment Cost Index, Q3 | 0.8% | 0.8% | Cooling, in line |
| Compensation, y/y | n/a | 3.4%, real down 0.4% | Wages losing to prices |
| Chicago PMI, October | ~55.8 | 55.3 | Expanding, decelerating |
| Euro area GDP, Q3 | ~0.2% | 0.4% | Double expectations |
The two year is at 4.60%, the lowest since mid September, and the ten year has fallen back below 5% to around 4.92% after trading above it earlier in the week. It is tempting to read that as relief. It is not. The Employment Cost Index decelerating to 0.8% with real compensation down 0.4% over the year, and Chicago PMI slipping from 58.8 to 55.3, is a rally driven by a growth scare rather than by any improvement in the inflation picture. Yesterday’s data had third quarter growth at 3.4% with core inflation stuck at 3.1%. Nothing published since changes that, and the committee that held on Wednesday with two dissents for a hike is not going to be moved by one soft regional survey.
What makes this uncomfortable is the company the bond rally is keeping. Crude is at a five month high, European natural gas is above 100 euros per megawatt hour and up roughly 42% this month, and euro area inflation is at 3.8%. A front end rallying on growth worries while energy runs is the classic late cycle configuration, and it is the one that leaves equities without a rescue from either direction. Duration buyers who stepped in this week are being paid for the right reason only if the slowdown is real, and the evidence for that is one PMI print.
Europe gave the better number. Euro area third quarter growth came in at 0.4%, roughly double expectations, with Germany up 0.2% and Spain contributing, while France contracted 0.1% and Italy stagnated. France also printed 3.8% inflation and is now the economy that keeps failing to grow while generating the fiscal spread, with the ten year gap over Germany back around 160 basis points. The European Central Bank held at 2.50% yesterday with an explicit warning on energy as an upside risk and a December hike as the base case, and Christine Lagarde leaves on February 7. A successor inheriting 3.8% inflation, an energy shock and a French fiscal problem is the live question for European duration, not the statement language.
The Bank of Japan held at 1.25% in an eight to one vote with Takata dissenting for an immediate increase, and will accelerate its exchange traded fund sales. Governor Ueda said underlying inflation has reached 2% and that the bank will continue to raise the policy rate, while declining to name a date and flagging currency, artificial intelligence and the Middle East as things he is watching. December remains priced above 80%. Japanese industrial production, employment and inflation all beat this morning. The yen firmed for a third session toward 157.2.
Energy and Geopolitics: The Premium Keeps Being Paid for Transit
Brent is up about 2.5% to $117.76 and WTI up 1.7% near $103.50, both at five month highs. Brent is up roughly 11% this month and WTI about 14%, a fourth consecutive monthly gain. Goldman estimates around $27 a barrel of geopolitical premium sitting on top of a fair value near $85, which is the cleanest quantification of the thing we have been arguing for weeks, namely that this is a freight and transit problem rather than a lost barrels problem.
Overnight added to it. The UK Maritime Trade Operations reported a tanker hit by explosive devices attached by a small boat in the Strait of Hormuz. Iran says tankers struck mines there, and disputes the American claim that key lanes have been cleared. The Houthis claimed a fresh wave of attacks on Saudi airports and oil facilities. Qatari mediators are back in Tehran and Vice President Vance is trying to reopen talks, which is why the market can still take a dollar out on a headline, but the direction of the preparation is the other way. Reporting says military options are being readied for after the November 3 midterms, Prime Minister Netanyahu is pushing for renewed strikes, and Secretary Hegseth fired Lieutenant General Brett Sylvia from the Joint Chiefs strategy and plans directorate over alleged leaks of classified material on Iran and Ukraine.
Our read is unchanged and is getting more asymmetric by the day. There are two trading sessions between here and an election after which the stated constraint on American action disappears. Nothing in the price today compensates you for that, because the market has spent eight months learning that nothing ever happens on any given morning. We would not be short energy into next week.
The Binance Breach Is Not a Crypto Story Only
Binance has been hit by roughly $6.3 billion of unauthorised outflows, the largest exchange breach on record, and has suspended deposits and withdrawals while it investigates. It says it has reserved more than $6.34 billion of company funds to cover the losses and that the unauthorised outflows have stopped, but investigators have found additional points of breach beyond the initially suspected compromised employee laptop. TRM Labs attributes the attack to North Korea. Tether and Circle have frozen somewhere between $1.8 billion and $2.4 billion of stablecoins.
Bitcoin fell below $77,000 for the first time in more than a month, touching $76,300, and roughly $1.5 billion of leveraged positions were liquidated in twenty four hours. The honest sequencing caveat is that the selloff was already running before the breach was public, so the hack accelerated something the equity market had started.
The reason this belongs in an equity note rather than a crypto one is that an exchange covering a $6.3 billion loss out of corporate funds is a solvency event that did not become a solvency event, and that distinction is worth exactly as much as you trust the disclosure. We have spent a month writing about collateral quality and funding structure in the neoclouds. This is the same category of question asked of a different balance sheet, arriving on the same morning that the two year rallies eight basis points. Risk appetite is being withdrawn from several places at once, and that is usually how these months end.
Elsewhere in Earnings
| Company | Result | Versus Street | Read |
|---|---|---|---|
| Chevron | Adj EPS $5.28, revenue $67.8bn | Beat ($4.89 est) | Shares up about 2% |
| Exxon | Adj EPS $4.23, revenue $113.2bn up 33% | Beat, dividend to $1.07, $20bn buyback | Shares up about 2% |
| Regeneron | Adj EPS $20.07, sales $5.1bn | Large beat ($16.19 and $4.5bn) | Shares up 7% |
| AbbVie | Adj EPS $3.82, FY raised to $13.95 to $13.99 | Beat and raise | Steady |
| Stryker | Guidance disappointed | Miss | Shares down 11% |
| Results and outlook disappointed | Miss | Shares down 20% |
The energy majors deserve more attention than they will get on a day like this. Exxon is now earning more from refining than from upstream for the first time in at least six years, which is a direct consequence of the effective closure of the Strait of Hormuz and the strikes on Saudi and Russian refining capacity. Across Exxon, Chevron, BP, Shell and TotalEnergies the sector produced about $56.3 billion of adjusted net income in the quarter. Both American majors chose balance sheet repair over a sharp increase in buybacks, which is the correct read of a windfall that depends on a war continuing.
The smaller moves in the pre-market tape tell the same rotation story as the indices. Alphatec down 33%, Reddit down 20%, Twilio and Roblox each down 18%, DXC down 12%, against Mondelez up 7%, Regeneron up 7%, Rivian up 6% and Gilead up 4%. Anything priced on growth multiples is being sold and anything priced on current cash is being bought, almost without reference to the quality of the individual print.
Asia and Europe
Asia was ugly outside China. The Nikkei fell 2.12% to 63,290.11 despite a batch of Japanese data that beat across industrial production, employment and inflation, the Topix lost 1.0%, Taiwan’s Taiex fell 2.5% and the Kospi dropped 3.48%. The CSI 300 fell 1.39% to 3,985.50. The exception is Hong Kong, where the Hang Seng rose about 1.5%, which continues a pattern we have noted several times this month in which Chinese equity trades on domestic policy and liquidity rather than on American technology earnings and therefore acts as a genuine diversifier in exactly the sessions when nothing else does.
Europe is holding up better than the futures implied. The Stoxx 600 is down only 0.21% at 647.74, the FTSE 100 is higher by about 0.2%, the DAX is down 0.4% and the Euro Stoxx 50 down 0.8%, with ASML off 4.5% carrying the semiconductor read-through and offset by Universal Music up as much as 15%, Sanofi up 5.8%, AstraZeneca up 3.3% and Puma up 7.6%. A technology shock that takes 2% out of Nasdaq futures and only 0.2% out of a European index is the same rotation visible at the regional level, and it is why European equity has quietly been the better place to be this month.
Also On The Tape
- Microsoft completed a $25 billion bond sale, which is the funding half of Wednesday’s story and worth holding against the fact that its shares rose on the print. Even the hyperscaler the market rewarded for restraint is in the debt market at scale
- Strategy’s chief financial officer Andrew Kang resigned on a day the company posted a record quarterly operating profit, and the shares fell. A finance chief leaving a levered balance sheet at a cycle turn is a datapoint, not a coincidence
- President Trump cancelled rallies in Kansas and Iowa ahead of Tuesday’s midterms, with prediction markets now assigning roughly 94% odds to Democrats taking the House. A divided Congress removes fiscal impulse from the 2027 outlook and matters more for rates than for equities
- The equal weight S&P rose 1.21% on Thursday while the Nasdaq 100 fell 1.38% and Alphabet had its largest one day decline on record at 12.64%. That single session is the month in miniature
- Next week brings Apple on Monday, the midterms on Tuesday, and the first window in which renewed American action against Iran becomes possible. Position sizing matters more than direction into that combination
On The Record
Yesterday we argued that capital spending had stopped being a demand signal and become a bill, and we wrote that the asymmetry into Amazon’s print sat in the capital spending line rather than in AWS growth. That was right, and it was right for the stated reason. AWS accelerated, the backlog grew, and the stock fell on $77.3 billion of capital and a $42.4 billion free cash flow outflow.
Where we were incomplete is Intel. We identified the memory shortage as an inflation input on Thursday through Samsung’s mobile loss and then failed to follow it into the companies that consume memory rather than produce it. Intel cut guidance partly because it cannot source data centre components, which makes the shortage a revenue problem rather than only a cost problem, and that is a different and more damaging mechanism than the financing argument we have been running all week. Both forces are now live at once, and we should have been carrying the second one with the same weight as the first.