US Market Recap: Intel's Guidance Collapses, the Chip Complex Gets Flattened, and AI Capex Starts Getting Audited

I. Three indexes: the gradient tells the whole story

The Dow fell only 1.09% while the Nasdaq lost 2.86%, a gap of nearly three to one, and the difference is composition. The Dow carries heavy weights in energy, consumer staples and financials — exactly the places money hid on Friday, with Chevron, Coca-Cola and JPMorgan absorbing part of the damage. The Nasdaq was held down by semiconductors and the AI-capex complex. At -7.2%, the SOX fell two and a half times as much as the Nasdaq: that was the epicenter.

The weekly picture is weak too, and it deteriorated at the tail end: the Dow -0.85%, the S&P 500 -0.57% and the Nasdaq -2.00% for the week. The Nasdaq was still climbing on Monday and Tuesday; Alphabet's report on Wednesday, then Intel and Amazon on Thursday and Friday, handed all of that back. In other words this was not a slow bleed. Two earnings events detonated within forty-eight hours.

II. Eleven sectors: six up, five down, and money backing out of duration

The spread between best and worst was 6.15 percentage points, narrower than Thursday's 9.1 points — but that does not mean the selling eased. Thursday's epicenter, communication services, fell only 1.11% on Friday, and the damage was spread across information technology instead. At -4.24%, tech alone accounted for the bulk of the index decline.

The six gainers share one trait: cash flow you can see today. Energy feeds on oil at 104 a barrel, staples on inelastic demand, and utilities and real estate on a 10-year yield back below 5%. Everything that fell, by contrast, is a story that pays off after 2027. This is a duration rotation, not a change in industry fundamentals. Industrials at -1.57% deserve a separate look: it has now underperformed for several sessions, which speaks to weakening real demand rather than valuation alone.

III. Mega-cap tech: six names down, none by more than 2%

Not one of the six fell more than 2%. That tells you the selling was not in the majors but in the supply chain behind them. Amazon is Friday's cleanest example: revenue and profit both missed, the stock traded down 6.3% premarket, and it still closed just -1.41% at 240.35 after touching 229.36 intraday. The market's read was explicit — weaker profit, but capex does not stop. That is a short-term pain for Amazon and bad news for everyone shipping into it.

Apple closed -1.23% at 344.63 with a 4.25% intraday range and a 334.35 low, the widest swing among the majors. Earnings are not until next week, so this looks less like fundamentals and more like a liquidity instrument that got used while money rotated between cash cows and the compute chain. Alphabet extended its decline by 1.40% but held the 280 area, which suggests the -12.64% session did not trigger a second wave of forced selling.

IV. Semiconductors: SOX -7.2%, and Intel's worst day since the 1970s

There is a contradiction buried in this table. Intel blamed memory prices for crushing PC demand, which by that logic should benefit Micron — yet Micron fell 5.30%. SanDisk fell 14.1% and Qualcomm 11.12%, neither in proportion to any PC exposure. The market is not selling fundamentals; it is selling the valuation multiple of the whole AI hardware chain. The closer a name sits to compute capex, the sooner it gets cut. The real fundamental differences will only show up in the next reporting round.

Intel itself is a different order of magnitude. It closed at 65.76, down 28.58%, with roughly $150 billion of market value gone and its worst session since the 1970s. This is not multiple compression; it is credit-event-level repricing, with share, process and capacity all questioned at once. Which is why -28.58% has to be read separately from peers at -5% to -11%: the first is a company problem, the rest is a lower valuation center, and the gap is an order of magnitude. Do not buy them as one trade.

V. Software, AI applications and China ADRs

The Nasdaq Golden Dragon China Index closed only slightly lower, about -0.3%, clearly better than the Nasdaq's -2.86%. The reason is simple: China ADRs are not part of the AI capex debate because their valuations never contained an option on monetizing compute. For two weeks this line has held the same pattern — US names fall, China ADRs hold up relatively. Alibaba losing 100 to close at 98.04, down 1.51%, is the ugliest print in the group, but it belongs to a different conversation than SOX -7.2%.

Oracle at -6.37% and Palantir at -6.27% were the direct casualties in software. Oracle's problem is not orders; it has to build the data centers before it can book the revenue, so higher rates and questioned capex hit it from both sides. Palantir is purer: at 184.11 the price embeds assumptions about many years out, and a higher discount rate compresses the multiple immediately.

VI. Six events that set the tape

• Intel: a solid third quarter, a fourth-quarter guide that broke (after the bell on October 29)

  Third-quarter revenue of $16.5 billion rose 21% and adjusted EPS of 0.52 beat the 0.43 consensus. Then the fourth-quarter guide landed at $15–16 billion against 170.7 billion expected, adjusted EPS of 0.25 and gross margin of 38%, with the CFO stating plainly that the company will not generate positive cash flow this year. Client segment revenue is guided down more than 20% sequentially, management sees PC weakness lasting into the first half of 2027, and the Israeli fab incident is expected to cost about $600 million of revenue across two quarters. The stock closed down 28.58% at 65.76.
• Amazon: a double miss on the quarter, capex unchanged at $220 billion

  Revenue and profit both came in below expectations, and the company held its 2026 capex plan at about $220 billion. The stock traded as low as -6.3% premarket and closed -1.41%. This matters more than Intel: it means falling profit will not cut AI spending, so the cash-flow strain on AI infrastructure keeps getting priced for several quarters.
• Alphabet fell 12.64% the day before, its worst session since listing

  Third-quarter revenue of $127.6 billion missed the 128.2 billion expected, cloud grew 103% to $30.7 billion, but free cash flow was -$9.1 billion for a second straight negative quarter, capex ran near $51.4 billion in the quarter, and the 2027 capex guide was pushed above $300 billion. Roughly $500 billion of market value vanished in a session, second only to Nvidia's January 2025 record. The stock extended the decline by 1.40% to 279.84 on Friday.
• Macro: inflation re-accelerating into a strong economy, December hike odds above 80%

  September PCE rose 0.5% month over month, up from 0.3%, and 3.6% year over year, up from 3.4%; third-quarter GDP grew 3.4% annualized, ahead of expectations. The Fed held rates at 3.75%–4.00% on October 28, but two voters favored a hike. CME FedWatch briefly put the odds of a 25 basis point hike in December above 80%. That is the common source of both the metals crash and the selling in long-duration assets.
• A fourth straight monthly gain for crude: Hormuz again

  Iran's navy said four tankers struck mines in the Strait of Hormuz, reviving supply-disruption risk. NYMEX December light crude rose 2.08% to 104.12 a barrel and ICE December Brent rose 2.40% to 117.64. Energy at +1.91% was therefore the only leadership group on Friday with a fundamental leg to stand on.
• Precious metals sold off hard: spot gold at a one-year low, silver below 50

  Spot gold touched 3,891 an ounce intraday, its lowest since October 2025, and is down more than 6% in October for a second consecutive monthly decline; spot silver broke 50 and is down more than 20% in October. Beyond rates, the margin calls and liquidity squeeze triggered by the tech selloff forced metals out as collateral — this time the haven bid never reached gold.

VII. Cross-asset: stocks down, gold down, oil up, bonds up

This grid carries the single most important message of the day: stocks down, gold down, oil up, bonds up, dollar down. Gold falling alongside equities while bonds rally alone means money is not seeking safety, it is seeking certain yield — a 4.94% ten-year coupon beats an AI narrative under audit. As long as December hike expectations do not reverse, this combination persists.

The dollar is the detail worth flagging. DXY at 101.89, down 0.14%, has now fallen three sessions in a row — with hike odds rising, a weaker dollar says this is not a dollar-funding squeeze but a rotation inside US equities. Gold down and the dollar down at the same time points to forced deleveraging in metals rather than a currency story.

VIII. Views

• The AI trade is not over, but the way it is priced has changed
  For two years the market paid a narrative premium for AI capex; now it wants a cash-flow timetable. Alphabet with two consecutive negative free-cash-flow quarters, Amazon spending $220 billion anyway and 2027 still stepping up — together they tell the market the spending peak is not here yet. What decides these stocks from now on is not order growth but who first specifies the year free cash flow turns positive.
• Intel is a company problem that exposed a cycle problem
  The guidance broke because of Intel's own issues — share, process, capacity — but it put the whole chain of weak PC demand plus memory costs squeezing bill-of-materials on the table, which is why AMD, Qualcomm and Micron were dragged in. Keep them apart: Intel at -28.58% is a credit-event repricing, peers at -5% to -11% are a lower valuation center. An order of magnitude separates them. Do not buy them as one trade.
• The haven bid skipped gold, which says rates are the dominant variable
  With December hike odds above 80% and real yields rising, carrying a non-yielding asset is expensive; the tech selloff then forced margin calls, and metals became the thing sold to raise cash. What actually caught the money was Treasuries, with the 10-year at 4.94% and -4bp. Until hike expectations reverse, every gold rally looks more like a technical repair than a trend.
• Two things to watch next week: whether oil holds 100, and what the majors say about capex
  If crude holds above 100, inflation expectations stay sticky, a December hike becomes easier to deliver, high-multiple growth keeps sliding, and energy's strength finally has duration. The other is whether a second company emerges late in earnings season that, like Amazon, misses on profit and keeps spending anyway. If it does, the de-rating in chips is not finished.