The S&P 500 rose 1.15% on the week and set a record on Tuesday, but without its usual engine: Technology finished last of eleven sectors after Thursday’s report that OpenAI’s revenue trails earlier signals broke the chip trade, and staples, utilities and energy did the lifting. Underneath, the Fed sharpened its case for more hikes while long-dated Treasuries found buyers at 24-year-high yields, and energy risk spread from the Strait of Hormuz to the platforms of the Gulf of Mexico. The market ends the week priced for higher rates and higher oil, and less sure who will pay for AI.

1. OpenAI Revenue Whiplash Splits the AI Trade: The FT’s $50 Billion Run-Rate Sinks Chips Thursday, Bloomberg’s $70 Billion Target Rescues Only Software Friday — Technology Finishes the Week Last

The core facts:The week opened with AI hardware at full strength. On Tuesday Marvell raised its fiscal 2028 revenue target to about $20 billion from about $18 billion and set a first fiscal 2031 goal of $70–90 billion; the stock rose 5.81%, Cisco 4.54% and Arista 4.09%, AMD closed at a record, and the S&P 500 made its first close above 7,800. On Thursday the Financial Times reported that OpenAI’s annualized revenue was near $50 billion at the end of September, “about $20bn less than has been previously signalled,” according to documents shared with investors. Technology fell 1.90% and the Nasdaq 100 1.39%, led by Oracle (−5.48%), Intel (−5.34%), SanDisk (−4.90%), Micron (−4.79%) and Broadcom (−4.35%). Late Thursday Bloomberg reported that OpenAI expects to reach or exceed $70 billion in annualized revenue by year-end and attributed the gap to a method that treats cloud-provider sales differently; OpenAI declined to comment. Friday’s rebound ran through software — Palantir +5.17%, Palo Alto Networks +5.09%, CrowdStrike +4.57%, Microsoft +2.38% — while Intel, AMD and Sandisk kept falling. The same day Nikkei reported Apple cut October iPhone 18 Pro component orders by at least 15% after a $100 price increase it blames on AI-driven memory costs.

Why it matters:OpenAI is the revenue line behind a large share of committed AI compute spending, so a dispute over its run-rate goes straight to who pays for the build-out — and Friday showed the market will not re-underwrite that spending on a disputed figure. The week therefore ended with the AI trade split in two: companies selling AI to users recovered, while the suppliers of capacity did not. The receipts are stark: Technology finished last of eleven sectors and the only one besides Industrials to fall (see sector rotation table in Section B), all five of the week’s biggest mega-cap decliners were chip names, and Palantir topped the gainers (see weekly movers table). Apple’s cut adds a second channel: AI demand is raising memory costs enough to dent volumes in consumer hardware, so part of the chipmakers’ pricing power is now coming out of their own customers.

What to watch:ASML’s results on Wednesday, October 14 and Taiwan Semiconductor’s on Thursday, October 15 are the first hard reads on AI capital spending this season; strong orders would test whether the chip group’s lag is fundamental or positioning. Any clarification from OpenAI itself would settle which run-rate is right.

↑ back to summary

2. Gulf Attacks Widen All Week — a Saudi Pipeline Hit, a Record Weekly Tanker Count, a Laden Supertanker Struck Off the UAE — as Trump Rules Out Iran Strikes Before the Midterms; Brent Holds Above $100 Every Session

The core facts:Monday: the Houthis claimed weekend strikes on Saudi Aramco sites, and a single Saudi source told AFP the East-West pipeline’s Khurais pumping station was hit again; up to four tanker attacks were logged near Hormuz from October 2 to 4. Crude still fell about 2% as Middle East exports topped pre-war levels and Aramco cut November prices for Asia to six-year lows. Tuesday: Saudi Arabia’s energy minister said the Hormuz-bypass pipeline was back to 5.8 million barrels a day, against capacity of about 7 million, while Houthi strikes hit airports in Jazan and Najran. Wednesday: a Reuters exclusive counted at least 12 attacks on tankers around Hormuz in the week to October 5, the most since the war began on February 28, and The Atlantic reported the White House had asked the Pentagon for Iran strike options usable before the November 3 midterms. Thursday: President Trump posted that the US “will not be attacking Iran at any time prior to the midterm elections,” and Treasury sanctioned 17 shadow-fleet tankers and 13 companies; only seven vessels had transited the strait on Tuesday. Friday: the fully laden supertanker GEM NO.2 was struck by an unknown projectile at anchor 13 nautical miles off the UAE, and Iran’s Revolutionary Guard navy said it had hit an LPG carrier on an “unauthorized” route south of the strait and that enforcement would no longer be confined to Hormuz. Brent finished at $104.25.

Why it matters:Two things moved in opposite directions. The most acute risk, a US strike, got a hard expiry date; the zone of attack spread beyond the strait into anchorages and loading areas that had been treated as safe, and Iran now claims that reach explicitly. Exports are recovering under rising fire rather than because fire has eased, which is why Brent never settled below $100 and finished the week higher even as Gulf exports climbed (see Energy table in Section B). Because crude was flat on Friday, the wider risk zone is not yet in the price; it reaches delivered costs through freight rates and war-risk insurance first. For rates, this is the energy channel the Fed minutes cited as one reason to keep hiking.

What to watch:Daily Hormuz transit counts and any US Central Command response to strikes outside the strait; the strike pledge expires on November 3, so the risk premium may rebuild into that date rather than fade.

↑ back to summary

3. The 10-Year Hits a 24-Year Closing High, Then Three Auctions Prove Buyers Exist at 5.3% — as the Fed Minutes and SpaceX’s $40 Billion Debt Plan Name AI Borrowing as the New Rival for Their Money

The core facts:On Monday the 10-year Treasury yield closed at 5.315% and the 30-year also posted its highest close in more than two decades, in a bear steepener that left the 2-year lower; the dollar reached an 18-month high as the euro fell below $1.12 on French bond-spread stress. Tuesday’s $58 billion 3-year auction cleared at 4.932%, the highest for the tenor since May 2006, on softer indirect demand. That evening the Financial Times reported SpaceX is seeking about $40 billion — roughly $10 billion of bank loans and $30 billion of investment-grade debt — to buy Nvidia chips, and Treasuries sold off Wednesday morning, the 10-year touching 5.364% intraday. The $39 billion 10-year auction then cleared at 5.300% with a 2.77 bid-to-cover and primary dealers left with just 2.5%. Wednesday’s FOMC minutes recorded market commentary attributing part of the rise in term premiums to “competition for capital from heavy private debt issuance to finance the development of artificial intelligence (AI) infrastructure.” Thursday’s $22 billion 30-year reopening cleared at 5.618% with dealers taking 6.8%. Freddie Mac’s 30-year mortgage rate rose to 7.40%.

Why it matters:The week answered the question hanging over the long end — whether there are end buyers at these yields — with yes, and the 10-year finished lower on the week after Monday proved the peak (see Vol & Treasuries table in Section B). But it answered a second question less comfortably: the price. Crowding out is showing up as a higher clearing yield, not as unsold bonds, and the Fed’s own minutes now name AI infrastructure borrowing as one of the competitors setting that price. SpaceX’s plan is a concrete instance, and it links this story directly to story #1: the AI build-out is simultaneously the market’s biggest source of equity earnings hopes and a growing claimant on the long-dated capital the Treasury needs. The cost lands on housing first, with mortgage rates more than a point above a year ago.

What to watch:The September Monthly Budget Statement on Tuesday, October 13, which closes the fiscal year and sizes the financing need, September CPI on Wednesday, October 14, and any formal SpaceX bond launch and its pricing against investment-grade peers.

↑ back to summary

4. Hurricane Isaias Shuts In 63% of US Gulf Oil Output Ahead of a Friday-Night Landfall; Crude Jumps 3% Thursday and Energy Finishes Among the Week’s Top Three Sectors

The core facts:Tropical Storm Isaias formed in the Gulf of Mexico on Wednesday, when operators had already shut in 25.08% of the Gulf’s oil production and 16.37% of its gas. By Thursday, as Isaias became the first hurricane of the 2026 Atlantic season, shut-ins reached 62.89% of oil and 57.35% of gas output, with 121 of 371 manned platforms evacuated; BP evacuated Na Kika and Thunder Horse, and Shell and Chevron curtailed operations. Landfall was forecast near the Mississippi-Alabama-Florida border on Friday night, with about 0.5 million barrels a day of refining capacity, including the Pascagoula refinery, in its path. Reuters estimates about 9 million barrels of production will be lost. WTI settled up $3.21 at $91.49 and Brent up $4.08 at $104.28 on Thursday, and Energy was the session’s best sector, with Chevron up 3.12% and ExxonMobil 2.71%. Crude was flat on Friday.

Why it matters:Offshore shut-ins usually reverse within days once platforms are reboarded, and Friday’s flat close says the market is pricing a temporary outage rather than damage. The lasting exposure is onshore: Gulf Coast refining, in a week when diesel was already the tightest part of the barrel and Washington was spending reserves and political capital to ease it (story #5). The storm also landed on the one sector that had been holding up the commodity side of the market: Energy finished the week up 3.23% and remains the year’s best sector (see sector rotation table in Section B), and Henry Hub gas was up nearly 5% on the week.

What to watch:Post-landfall damage reports from Gulf Coast refineries over the weekend and the pace of platform reboarding; WTI holding above $90 into next week would signal the market is pricing damage rather than a short outage.

↑ back to summary

5. Washington Pulls Three Levers on the Diesel Shortage in Five Days: A Tax-Relief Order, an IEA Release Speed-Up With Diesel First, and a License for Russian Diesel Imports

The core facts:Monday: President Trump signed an executive order in Grand Island, Nebraska temporarily authorizing highway use of tax-free dyed diesel and deferring the 24.4-cent federal diesel excise tax through the end of 2026; the White House put the saving at about $60 per 250-gallon fill. Diesel reached a record of about $6.50 a gallon last month, and Energy Department data showed the Strategic Petroleum Reserve at 283 million barrels, its lowest since October 1982. Wednesday: IEA member governments backed accelerating the releases announced in March and prioritising diesel “given the current tightness in diesel markets”; about 325 million barrels have been released, with about 100 million pledged but not yet delivered. Crude settled lower despite a surprise 3.2 million-barrel US inventory draw. Friday: Treasury issued General License 135, authorizing the importation, including into the United States, of Russian-origin diesel until April 7, 2027, after the President said Russia would “immediately supply over 300,000 Tons” with more to follow in November; Russia’s own ban on most diesel exports runs through October 31.

Why it matters:Diesel, not crude, is where the energy shock bites US freight, farm and goods costs, and this week the administration attacked it from the tax side, the stockpile side and the sanctions side at once. The Russian license is the significant one: it addresses the product shortage directly and reverses sanctions policy toward Moscow less than four weeks before the midterms, carrying political risk alongside its disinflationary effect. The limits are visible too. The tax measure is a deferral, the IEA step brings forward barrels rather than adding them, and a reserve at a 44-year low leaves less cushion for the wider attacks in story #2 or storm damage from story #4.

What to watch:Whether the first 300,000 tons of Russian diesel actually load, next week’s IEA Governing Board review, and Russia’s decision on its export ban when it expires on October 31; September CPI on October 14 shows how much energy is adding to inflation before any relief arrives.

↑ back to summary

6. The White House Builds a Process to Fire Fed Governor Lisa Cook, With a Hearing on November 5, Two Days After Trump Says the Board “Would Like to See the Country Do Badly”

The core facts:On Wednesday, asked about mortgage rates in the Oval Office, President Trump called Fed Chair Kevin Warsh “great” but said the rest of the Board “would like to see the country do badly, in my opinion, because I think interest rates should come down.” Treasury Secretary Bessent said inflation, mortgage rates “and the 10-year will come back down” once the Iran conflict ends. On Friday a presidential memorandum dated October 7 was made public establishing a three-member committee — the Assistant to the President for Economic Policy, the chair of the EEOC and the director of the Office of Government Ethics — to report “whether there is ‘cause’ for Governor Cook’s removal” under 12 U.S.C. 242 as construed by the Supreme Court in Trump v. Cook. Cook’s written position is due three days before a closed, four-hour hearing at the White House on Thursday, November 5, with a post-hearing statement due by November 10. No market reaction was identified.

Why it matters:The Supreme Court allowed a renewed removal attempt if it came with notice and a hearing; this memorandum supplies exactly that, turning a legal tail risk into a scheduled process with a decision plausible before year-end. Read together with Wednesday’s remarks, the pressure is aimed at the governors rather than the Chair, at the moment most of the Committee expects another hike. The hearing sits between the October 27–28 meeting and December’s, so the composition of the Board that decides the next move is now itself uncertain. A removal through an executive fact-finding process would make every Board seat more exposed to politics, and that is a risk investors would price into long-dated yields — already a few basis points from a 24-year high (story #3).

What to watch:Cook’s written response due Monday, November 2 and any legal challenge to the committee before the November 5 hearing; a widening gap between 10-year and 2-year yields would be the market’s signal of a rising term premium.

↑ back to summary

7. Google Underwrites 890 MW of New Nuclear Capacity From Constellation for 20 Years; Constellation Jumps 12% and Utilities Hold Their Gain Through the Chip Selloff

The core facts:On Tuesday Google and Constellation Energy announced a 20-year power purchase agreement to bring 890 megawatts of new nuclear capacity onto the PJM grid through uprates at 11 Constellation units in Illinois, Pennsylvania and New Jersey, plus a 15-year supply agreement for a further 2,700 MW from Constellation’s existing PJM fleet. Constellation will invest more than $4.3 billion, with the first uprate expected by 2028, and selected Google Cloud and Gemini Enterprise under an expanded five-year technology alliance. Constellation shares rose 12.25% to $300.40 and Utilities led all eleven sectors that day with a 2.79% gain.

Why it matters:A hyperscaler signing for two decades turns AI power demand into contracted utility revenue rather than a forecast, and uprating existing reactors is the fastest route to new firm, carbon-free capacity. The weekly view adds the more useful signal: Utilities held the gain through Thursday’s AI selloff and finished the week second among sectors (see sector rotation table in Section B), while the chipmakers in story #1 gave theirs back. The power leg of the AI build-out, backed by long contracts, traded as the safer way to own the theme in a week the compute leg was questioned. It lands on a sector still down nearly 12% over six months, so the re-rating has room if more deals follow.

What to watch:Whether other hyperscalers sign comparable uprate or supply contracts, and whether Utilities extend their leadership once the defensive rotation of this week fades.

↑ back to summary

8. SpaceX Buys Nationwide 800 MHz Spectrum for Starlink Mobile, and AT&T, T-Mobile and Verizon Fall 5% to 7% After Hours

The core facts:After Thursday’s close SpaceX agreed to acquire all of Grain Management’s nationwide 800 MHz spectrum, up to 14 megahertz of paired low-band frequencies, for its Starlink Mobile service. Terms were not disclosed and the deal needs FCC approval; Grain bought the licenses from T-Mobile for $2.9 billion and was reported in August to be seeking about $6 billion. Low-band spectrum lets Starlink Mobile add a terrestrial layer that reaches indoors, and most existing phones already support the band. Earlier in the week the FCC approved Starlink Mobile’s second-generation constellation of 15,000 satellites. In after-hours trading AT&T fell 6.75%, T-Mobile 5.4% and Verizon 5%, and on Friday Communication Services was one of only two sectors to close lower as the carriers sank.

Why it matters:The deal turns SpaceX from a satellite partner of the carriers into a would-be carrier with the low-band spectrum that indoor coverage requires. The three national networks are owned as defensive income stocks, and a credible new competitor goes to the stability of the subscriber base that justifies that valuation — in a week when investors were otherwise rotating into exactly that kind of defensive income. The challenger is also one of the AI build-out’s biggest new borrowers (story #3), so its expansion into telecom will be financed in the same bond market the carriers depend on.

What to watch:Any response from AT&T, Verizon or T-Mobile, the FCC review timetable for the transfer, and carrier commentary on subscriber competition in third-quarter results.

↑ back to summary

Fed Minutes Show Most Officials See Another Hike by Year-End, and Waller and Musalem Back It — With Flexibility on Timing (Federal Reserve, Wed, Oct 7 – Thu, Oct 8)

What they’re saying:Minutes of the September 15–16 meeting, which raised the target range a quarter point to 3.75%–4.00% on a 12-0 vote, state that “most participants assessed that another increase in the target range for the federal funds rate would likely be appropriate by year end,” with inflation risk “skewed to the upside” and several officials calling policy “not restrictive or only mildly restrictive.” On Thursday Governor Waller said he anticipates additional hikes but that they “do not need to come at consecutive meetings,” citing core PCE inflation of 3% and noting that 16 of 18 September projections showed at least one more increase this year. St. Louis Fed President Musalem said “more monetary policy firming will be required,” with rates going up “in the next six to nine months,” but declined to prejudge October. On Tuesday San Francisco Fed President Daly said further hikes depend on whether the tariff, oil and AI shocks fade or compound.

The context:The news in the week’s Fed messaging was timing, not direction. Waller’s “not at consecutive meetings” and Musalem’s six-to-nine-month window stretch the cycle into 2027 rather than front-loading it, which is why the market took three hawkish signals in stride: the 2-year and 10-year finished the week 3.6 and 3.2 bps lower (see Vol & Treasuries table in Section B) and Polymarket’s 2026 hike odds were unchanged at 100% (see Polymarket table above). The minutes’ split rationale matters for what comes next: officials citing energy shocks could lose their case if fuel prices fall, while those citing AI-driven demand would not. On Friday futures put the odds of a December increase at 82%.

What to watch:September CPI on Wednesday, October 14, the last inflation report before the October 27–28 meeting; a hot reading would test how much timing flexibility survives.

Inflation Expectations Rise on Two Surveys — NY Fed One-Year at 3.9%, Highest Since May 2023; Michigan Long-Run at 3.5% as Sentiment Misses at 46.3 (NY Fed, Wed, Oct 7; University of Michigan, Fri, Oct 9)

What they’re saying:The New York Fed’s September Survey of Consumer Expectations put median one-year inflation expectations at 3.9%, up from 3.6%, with three-year expectations at 3.3% and five-year steady at 3.0%; the perceived probability of losing one’s job fell to 13.5%, the lowest since December 2024. Michigan’s preliminary October sentiment index fell to 46.3 from 48.1, below the 47.6 consensus, as current conditions dropped to 44.7 from 50.9 against a 50.5 forecast while expectations rose to 47.3. Year-ahead inflation expectations rose to 4.7% from 4.6% and long-run expectations to 3.5% from 3.4%, both the highest since May. A Michigan special report found only 31% of consumers plan to keep buying as usual where prices have jumped; 54% plan to cut back and 16% to stop.

The context:Two independent surveys moved the same way in the same week, and it was the long-run Michigan reading, now up two months running, that matters most to a Fed trying to show the energy shock is not becoming entrenched. Both surveys also show households feeling secure in their jobs while expecting to cut back where prices have jumped, so the pressure is landing on what they buy rather than on whether they work — the version of the problem that shows up in spending before it shows up in payrolls. The 2-year yield rose 3.7 bps on Friday, the only day it rose all week, though no source tied the move directly to the release.

What to watch:September CPI on Wednesday, October 14 (expected 3.6% from 3.4%), September retail sales on Thursday, October 15 as the first hard test of whether weaker sentiment is reaching spending, and the final October Michigan reading on Friday, October 23.

ISM Services Holds at 54.9, in Line With Forecasts, but Prices Paid Jumps to 74.0, Highest Since July 2022; S&P Global and Freight Gauges Confirm the Cost Pressure (ISM, Mon, Oct 5)

What they’re saying:The ISM Services PMI registered 54.9 in September against a 55.0 consensus and 55.4 in August, its 27th straight month of expansion. Business activity fell 5.2 points to 56.5 and new orders slipped to 59.8, while employment rose 2.3 points to 50.1. The Prices Index rose 1.4 points to 74.0, above 70 for the sixth time in seven months. S&P Global’s final services PMI rose to 58.8, the fastest expansion since July 2021, with input cost inflation at its steepest since November 2022. On Tuesday the Logistics Managers’ Index rose to 70.2, with transportation prices at 92.7 and capacity contracting.

The context:Monday’s market ignored it — the 2-year fell as hike odds stayed low after the September payrolls miss — but the print set the agenda the Fed then spoke to for the rest of the week. Services is where most core inflation sits, and three separate gauges now show costs rising with capacity tight, the combination that passes increases through to final prices. It is the clearest evidence of the week that the price pressure is not just energy, which weakens the argument that a fall in oil alone would let the Fed stop.

What to watch:Core CPI on Wednesday, October 14 (expected 0.2% on the month) for whether services prices are reaching consumers, and September PPI on Thursday, October 15.

Jobless Claims Fall to 197,000, Near the Lowest Since 1969, Even as the Conference Board’s Employment Trends Index Drops 0.5% (Department of Labor, Thu, Oct 8; Conference Board, Mon, Oct 5)

What they’re saying:Initial claims fell 2,000 to 197,000 in the week ending October 3, below the 200,000 expected, and the four-week average fell to 198,000; continuing claims rose 17,000 to 1,716,000. The Conference Board’s Employment Trends Index fell to 107.56 in September from 108.08, ending two monthly gains, as the share of consumers saying jobs are “hard to get” rose to 21.9% from 20.3% and the share of small firms with positions they could not fill fell to 32% from 35%.

The context:The week’s labor data describe a market that has stopped hiring without starting to fire. Claims near 1969 lows say layoffs are not rising; the trends index, the consumer “hard to get” reading and the September payrolls gain of 29,000 reported the previous Friday say hiring has slowed. That combination gave the Fed no reason to soften: Waller called the labor market “solid and stable” the day the claims printed, and the recession odds in the Polymarket table edged down a point. The 17,000 rise in continuing claims is the detail to track, because it is where slower hiring shows up first.

What to watch:Initial claims on Thursday, October 15 (expected 197,000) and the ADP weekly employment estimate on Tuesday, October 13.

GDPNow Holds Near 3.6% for Q3 Despite a Wider-Than-Expected $105.6 Billion Trade Deficit — but Inventories Supply More Than Half of It (BEA, Tue, Oct 6; Atlanta Fed, Thu, Oct 8)

What they’re saying:The August goods and services deficit widened to $105.6 billion from a revised $92.8 billion, wider than the $102 billion consensus, as imports rose 4.3% on crude oil, gold and semiconductors. The Atlanta Fed’s GDPNow model cut net exports’ contribution to −2.69 points but held third-quarter growth at 3.7% on stronger consumption and investment nowcasts. On Thursday, after a downward revision to August wholesale inventories, it eased to 3.6%, with inventory investment contributing 1.98 points. Wholesale sales rose 1.8% in August, while consumer credit rose only $8.3 billion against $15 billion expected as card balances contracted.

The context:There is no growth scare to offset the inflation side of the ledger: a nowcast near 3.6% even after a 4.3% import surge supports the Fed’s view that activity is firming, which is why the week’s recession odds barely moved. The quality is the concern. With stockbuilding supplying about 55% of the estimate and household credit growth slowing, the quarter rests more on inventories than on final demand — and inventory gains tend to reverse.

What to watch:September retail sales, August business inventories and the GDPNow update, all on Thursday, October 15.

1. PepsiCo (PEP): +3.73% on Thursday | Beats on Revenue and Core EPS, but Cuts Full-Year Core EPS Growth to 2.5%–3.5% to Fund a North America Reset

The Numbers:Reported before the open on Thursday, October 8. Revenue $25.27 billion vs. $24.95 billion expected (+5.6% year over year; organic +3.1%). Core EPS $2.34 vs. $2.29 expected (+2%); reported EPS $2.23 (+17%). Fiscal 2026 guidance: organic revenue growth about 3% (prior 2%–4%), net revenue growth about 6% (prior 4%–6%), core EPS growth 2.5%–3.5% (prior: low end of 5%–7%), and core constant-currency EPS growth 1%–2% (prior: low end of 4%–6%).

The Problem/Win:The beat came from abroad and from one-offs. Organic revenue grew 9% in EMEA and Asia Pacific Foods, 7% in International Beverages and 6% in Latin America Foods, while both North American businesses were flat and North American beverage volume fell 3%. Core operating profit rose 3% despite 35 basis points of margin contraction, and a 4-point favorable tariff-refund effect was part of that gain. Management is lowering the year’s earnings growth to spend more on North American innovation, funded by “additional structural cost reduction actions.”

The Ripple:Consumer Defensive rose 2.17% on Thursday, with Coca-Cola up 2.27% and Philip Morris up 4.05%, though no source tied those moves to PepsiCo’s report. The sector went on to lead all eleven for the week.

What It Means:Investors paid for international volume and a reinvestment story while accepting a sharp cut to this year’s earnings growth — and the tariff refund that flattered the quarter is a non-recurring item that a nationwide class of importers is now also pursuing. Until North American organic growth turns, the multiple rests on the promise of the reset rather than on its results.

What to watch:The size and timing of the structural cost actions, and whether North American organic revenue turns positive next quarter.

Q3 2026 earnings season is just beginning (3.8% of the S&P 500 reported as of October 9). The large banks, Johnson & Johnson and UnitedHealth open the season in earnest on Tuesday, October 13; ASML, Bank of America, Morgan Stanley, BlackRock and Progressive follow on Wednesday, October 14; and Charles Schwab and Prologis report on Thursday, October 15.

JPMorgan Chase (JPM) — BMO, Tuesday, October 13 — Consensus EPS $5.93 on revenue of $51.23 billion. Key focus: net interest income guidance with Fed officials signaling more hikes and the 2-year yield at 4.793%, trading revenue in a volatile rates quarter, and credit commentary after its own strategists flagged the most deeply distressed leveraged loans since March 2020.

Johnson & Johnson (JNJ) — BMO, Tuesday, October 13 — Consensus EPS $2.50 on revenue of $25.38 billion. Key focus: medtech exposure to the coming Section 232 device tariffs after Becton Dickinson’s investment-for-relief deal, and pharmaceutical growth guidance for 2027.

UnitedHealth Group (UNH) — BMO, Tuesday, October 13 — Consensus EPS $4.15 on revenue of $111.31 billion. Key focus: medical cost trends and margin guidance for 2027, plus its first comments on the 2027 Medicare Advantage Star Ratings published October 8, which drive quality-bonus revenue.

Goldman Sachs (GS) — BMO, Tuesday, October 13 — Consensus EPS $12.86 on revenue of $16.83 billion. Key focus: advisory and underwriting fees in an active quarter for large mergers and AI-related debt financing, and fixed-income trading revenue with the 10-year above 5.2%.

Wells Fargo (WFC) — BMO, Tuesday, October 13 — Consensus EPS $1.85 on revenue of $22.31 billion. Key focus: management’s first comments on HUD’s new fair-lending investigation, mortgage volumes with Freddie Mac’s 30-year rate at 7.40%, and net interest income guidance.

Citigroup (C) — BMO, Tuesday, October 13 — Consensus EPS $2.65 on revenue of $23.72 billion. Key focus: markets and services revenue, card credit costs, and progress on its return-on-equity targets.

ASML Holding (ASML) — BMO, Wednesday, October 14 — Consensus EPS $12.14 on revenue of $13.16 billion. Key focus: EUV order intake from AI and memory customers after a week in which chipmakers fell on doubts over OpenAI’s revenue and did not join Friday’s software rebound; the outlook for China sales; and 2027 guidance as the first major read on chip-equipment demand this season.

Bank of America (BAC) — BMO, Wednesday, October 14 — Consensus EPS $1.10 on revenue of $30.58 billion. Key focus: net interest income guidance and deposit costs as rates rise, consumer credit quality as Michigan’s current-conditions gauge falls to 44.7, and trading revenue.

Morgan Stanley (MS) — BMO, Wednesday, October 14 — Consensus EPS $2.93 on revenue of $19.93 billion. Key focus: wealth-management net new assets and fee growth, and investment-banking and equities trading revenue.

BlackRock (BLK) — BMO, Wednesday, October 14 — Consensus EPS $14.33 on revenue of $7.43 billion. Key focus: fixed-income fund flows with long-dated yields above 5%, private-markets fundraising, and the base-fee effect of higher equity markets.

Progressive (PGR) — BMO, Wednesday, October 14 — Consensus EPS $4.30 on revenue of $22.61 billion. Key focus: underwriting margins after the August combined ratio worsened to 89.3% from 83.1% a year earlier as premium growth slowed to 6%, and auto policy growth.

Charles Schwab (SCHW) — BMO, Thursday, October 15 — Consensus EPS $1.66 on revenue of $7.20 billion. Key focus: net interest revenue and client cash balances as the Fed signals more hikes, trading activity in a volatile quarter, and net new client assets.

Prologis (PLD) — BMO, Thursday, October 15 — Consensus EPS $0.78 on revenue of $2.18 billion. Key focus: warehouse leasing demand and occupancy as tariff refunds and higher rates reshape importers’ inventory decisions, rent growth on renewals, and progress on its data-center development pipeline.

No company above $100 billion reports on Monday, October 12 (Columbus Day), or Friday, October 16. Taiwan Semiconductor (ADR) reports on Thursday, October 15, and BNY Mellon ($96.83 billion) reports the same morning, just below this section’s threshold.

1. Does Wednesday’s CPI turn “by year-end” into “October 28”? The minutes, Waller and Musalem all point to another hike but leave the meeting open; headline is expected at 3.6% and core at 0.2% on the month, so a hot core reading is what would pull the move forward.

2. Can ASML and TSMC repair the chip trade the OpenAI revenue dispute broke? ASML reports Wednesday and Taiwan Semiconductor Thursday; strong AI orders would say this week’s chip selloff was positioning, while caution would confirm that the market is right to separate AI’s users from its suppliers.

3. Do the banks confirm the credit warnings building beneath the index? JPMorgan, Goldman Sachs, Wells Fargo and Citigroup report Tuesday, after JPMorgan’s own strategists counted $65 billion of deeply distressed leveraged loans and Wells Fargo drew a HUD fair-lending probe; with yields this high, net interest income guidance and loan-loss provisions will show who is benefiting and who is carrying the risk.

4. Does Isaias leave lasting damage — and does Russian diesel actually arrive? Post-landfall refinery reports and Thursday’s EIA data will show whether the storm cost barrels or only days, while the first loading of the 300,000 tons of Russian diesel tests whether this week’s supply levers deliver real relief.

5. Are consumers acting on their pessimism yet? Michigan’s current-conditions index fell 12% and most respondents plan to cut back where prices jumped; Thursday’s retail sales will show whether intentions have turned into spending cuts, which would change the growth side of the Fed’s calculation.

WHY THIS CHARTOf the week’s five charts, Thursday’s is the only one that sits where the two biggest stories meet: the Treasury auction it dissects cleared on the same afternoon the Fed minutes named AI infrastructure borrowing as a rival for long-term capital, the day after SpaceX’s $40 billion debt plan surfaced. Monday’s job-loss chart and Friday’s household-expectations chart each explain one input to the Fed’s next move; this one explains why the cost of financing the AI build-out and the cost of financing the government have become the same question.

ORIGINAL CHART ANALYSIS — FROM THURSDAY’S MIBYields this high are usually read as a buyers’ strike, but Wednesday’s 10-year auction shows buyers setting the price. With the amount on offer fixed, a shortage of buyers shows up in one of two ways: the primary dealers — banks obliged to bid at every Treasury auction — keep the leftovers, or the yield rises until investors take nearly the lot. Wednesday was the second: dealers were left the smallest share of 221 such sales since 2008, and $2.77 was bid per dollar sold, matching the highest since January 2016. Treasury fills bids from the lowest yield up, and the last one accepted sets every winner’s yield — so 5.3% was the price investors themselves named. Fed minutes released an hour later suggest why: they cite market commentary naming one rival for investors’ money — heavy private borrowing for AI infrastructure, lifting term premiums, the extra yield demanded to lend for years. SpaceX’s reported $40bn borrowing talks are the latest example. That is crowding out — a higher price, not unsold bonds — and it is no one-off: the three smallest 10-year dealer shares since 2008 all came in the last 13 months. Thursday’s 30-year is the edge to watch: at 5.618%, its highest auction yield since 1999, it left dealers 6.79% — still the third-smallest since 2008, but nearly three times the 10-year’s, on a bond that loses almost twice as much for the same rise in yields. Treasury will not run out of buyers — only out of cheap ones. What it means: For homebuyers, cheaper mortgages will need investors to accept lower yields — buyers are not what is missing. Freddie Mac’s 30-year rate hit 7.40% today, the highest since November 2023, from about 6% in February. For stock investors, companies borrowing to fund AI pay a premium over that 5.3%, so their build-out will get dearer as yields rise. For the economy, Treasury now borrows for ten years at 5.3% against an average 3.4% on its notes, so its interest bill will keep climbing. What would change this: a 10-year sale leaving dealers over 10%, last seen in May.