Closing Briefing — Monday, August 3, 2026

Sources: CNBC market live blog (“Dow closes at record high”; ISM, Iran, Amazon, Marriott, Alibaba, Asia and Europe items); Bloomberg Markets Wrap (“Stocks Close Near Record as Amazon Tops $3 Trillion Market Cap”), the Bloomberg /markets quote board, “Oil Slips as Trump Says Iran, US in Talks Over Hormuz Reopening” (official WTI/Brent settlements) and “Dip Buyers Need to Heed the Market’s Warning Signs”; WSJ Markets (“U.S. Stocks Rally, While Oil Slides on Middle East Diplomacy Hopes”; “The Buffett Indicator Keeps Flashing Red. Is It Broken?”; What’s News: “Iran Says No Talks With the U.S. on Hormuz Strait”); Investing.com Major Indices, NDX, SOX, Dow-component, S&P-500-component and trending-stock boards (all stamped 15:59:59); U.S. Treasury Text View; CME FedWatch and Investing.com Fed Rate Monitor; FRED ICE BofA OAS series and NY Fed reference rates; TradingEconomics commodities and currencies boards; Finviz Groups (Performance view, rendered); NY Fed August calendar; Earnings Whispers day pages. Full link list in the companion DataNotes file.

1. The whole session is one trade — the Hormuz risk premium came out of crude — and the most useful thing on the tape is that the people closest to the barrels do not believe it. WTI September settled −5.1% at $80.34, Brent October −4.7% at $83.77 (Bloomberg), after Trump called off a planned attack and said US–Iran talks on reopening the strait were live “literally by tomorrow.” Within hours Iran’s foreign ministry said there were no immediate plans for peace talks and, per WSJ’s What’s News, “Iran Says No Talks With the U.S. on Hormuz Strait.” TD Securities’ Ryan McKay: “The market read for crude oil is overly bearish, as to be expected with the asymmetric downside sensitivity to headlines, but we anticipate the market may be quickly disappointed when the flows don’t surge like they did post MoU.” UK Maritime Trade Operations reported a tanker off Oman experienced an explosion in close proximity on Sunday. Commonwealth Bank estimates Hormuz traffic at 30–35% of pre-war levels, with 50–60% enough to reassert oversupply. Expression: sell the de-escalation rally in crude only through options, not futures — a $26 monthly range is not a market to be short outright. Invalidation: a verified convoy or an Omani-brokered transit corridor actually operating.

2. Breadth flipped completely in one session, and the Russell beating the S&P is the cleanest confirmation available that Friday’s narrowness was a positioning artefact. Russell 2000 +1.72% versus S&P 500 +1.48% and Nasdaq 100 +1.78%; ten of eleven Finviz groups green versus six of eleven red on Friday. Communication Services +3.82%, Industrials +2.45%, Consumer Cyclical +2.35%, Technology +1.58% — and critically, Technology was green this time, having been down 0.49% on Friday’s 1.00% Nasdaq day. JPMorgan’s de-grossing note is the supporting evidence: hedge funds “have largely completed” the step back from semis, and “the collapse of momentum signals in tech-heavy equity indexes is a sign that momentum traders such as CTAs have largely unwound previous long equity positions in the Nasdaq, Kospi, Taiwan and Nikkei.” A market with the sellers gone and the shorts on rallies below a 5% crude drop.

3. The SOX round-tripped Friday’s fade in the opposite direction — it traded −3.44% intraday and closed +1.06%. The prints: SOX low 10,922.3, high 11,495.3, close 11,430.4 (+119.3); Micron low $770.10 (−6.4%) and close $829.50 (+0.79%); SanDisk low $1,121.27 (−7.7%) and close $1,288.58 (+6.07%); AMD low $455.30 (−4.4%) and close $484.46 (+1.75%); CoreWeave low $69.89 (−2.6%) and close $85.74 (+19.47%); IREN +8.02%; Marvell +3.34%; Broadcom +0.76%. Friday the same complex printed new intraday highs and closed at the lows; Monday it printed new lows and closed at the highs. Two consecutive full reversals in opposite directions is the signature of a large basket being worked, not of a fundamental re-rating in either direction. Expression: the dispersion trade beats the direction trade — own AMD/NVDA against MU/SNDK into the prints, and buy the straddle rather than picking a side. Catalyst: AMD Tuesday 8/4 4:15 p.m. ET, then WDC and SNDK Wednesday 8/5.

4. The bond market delivered a bull-flattening from the 2-year out and an 8 bp bill cheapening at the very front — and the two have nothing to do with each other. Official par: 3M +8 bp to 3.91%, but 2Y −3 to 4.25%, 3Y −2 to 4.32%, 5Y −5 to 4.40%, 7Y −5 to 4.54%, 10Y −5 to 4.70%, 20Y −5 to 5.23%, 30Y −4 to 5.23%. 3M10Y flattened 13 bp to +79 bp; 2s10s flattened 2 bp to +45 bp; 2s30s flattened 1 bp to +98 bp. The belly-and-long-end rally is the oil trade — five basis points of inflation compensation coming out on a 5% crude drop. The 3-month is a different animal: bills cheapened into the quarterly refunding window while the coupon curve rallied, a financing signal, not a policy signal (§9: SOFR at 3.66%, one basis point above IORB, on a record $3.205tn of volume). Level to watch: 10Y 4.60%, below which the equity multiple genuinely re-rates; 30Y 5.35% remains the upside pain point.

5. Bristol Myers spiked 5%+ on a $400bn merger headline and closed +0.21%. That is the most instructive fade of the session. The FT reported AstraZeneca and Bristol Myers Squibb have been in merger talks for several months, a combination that would rank as the largest ever in pharma. BMY closed at $65.45, +0.21%, having traded as high as $68.10 — a 3.9% intraday give-back on 24.6m shares. AstraZeneca’s ADR fell as much as 7%, dropping to the bottom of the Stoxx 600 and wiping out the FTSE 100’s morning gain. The State Street Health Care Select Sector SPDR was up more than 1% pre-market, on pace to eclipse its $163.74 record close, and ended the day little changed (CNBC). Jefferies: “Given the strength of AZ’s growth and innovation profile, we are a bit perplexed… if there is one company that doesn’t need financial engineering, it’s AZ.” Healthcare closed −0.34%. The market prices this as value-destroying for the acquirer and low-probability for the target — a sell signal on merger-arb spreads, not a buy signal on either equity.

6. Amazon crossed $3 trillion, and the striking part is how little it cost the rest of the tape this time. AMZN +4.59% to $284.04 on 79.3m shares, following +15.31% on Friday and +17% on the week, its biggest weekly gain since 2015. Only the fifth company ever above $3tn. On Friday, Amazon’s blowout bought the index 52 points and the rest of the market nothing. Monday, Amazon added 4.6% and Alphabet 4.88%, Meta 6.01%, Microsoft 4.94% and Oracle 9.23% alongside it — the megacap bid generalised. That is what a genuine re-risking looks like versus a squeeze, and it is the strongest single argument that the July drawdown is over. The counter-argument is §13’s: BofA reports 77% of S&P 500 companies beating on EPS, the highest since 2021, with Q2 earnings tracking +45% y/y including Alphabet and Amazon mark-ups — so the bar for the rest of the year is now extraordinarily high.

7. Boeing’s 737 MAX 7 certification is the largest single-name regulatory event of the year for the industrial complex. BA +8.01% to $233.45 on 10.0m shares, range $219.21–$234.09. The FAA said the approval “reflects years of sustained work to resolve complex technical issues and complete a thorough review of the airplane’s design and supporting safety analyses,” ending nearly a decade of delay driven by the MAX 8 crashes, subsequent manufacturing crises and the engine anti-icing redesign. Southwest is the launch customer. Roughly 80% of covering analysts now rate Boeing the equivalent of buy against an S&P average near 55–60%, with a mean target around $276. The read-across runs to the supplier chain and to Industrials +2.45%, the third-best group of the day.

8. The analyst tape was unusually consequential and produced two clean second-order tells. Upgrades: Corning to Buy from Hold at Truist, target $175 (cut from $205) — GLW +6.07% to $146.64; Eaton to Outperform from In Line at Evercore ISI, target $502 from $453 — ETN +5.55% to $438.23; Ingersoll Rand to Buy from Hold at Stifel, target $101 from $84 — IR +5.28% to $87.78; LyondellBasell to Overweight from Neutral at JPMorgan, target $80 from $75; CoreWeave to Buy at Truist. Downgrades: eBay to Underweight from Equal Weight at Wells Fargo — EBAY −6.03% to $107.13, the worst S&P performer after Marriott; NXP Semiconductors to Neutral from Buy at UBS (Francois-Xavier Bouvignies), target $270 from $305 — NXPI −2.18%. The tell: Starbucks was upgraded to Hold from Sell at Melius and still closed −1.79% at $103.37 on a day the S&P rose 1.48% — a stock that cannot rally on an upgrade in a broad tape has a demand problem the sell side has not yet marked.

9. Energy was the only red sector and the refiners were worse than the integrateds — the crack-spread trade unwound faster than the barrel. Energy −1.26% with Marathon Petroleum −2.98%, Occidental −2.80%, Devon −1.21%, Diamondback −2.07%, EOG −2.02%, Valero −1.71%, ONEOK −1.67%, Williams −1.51%, ConocoPhillips −1.10%, Chevron −1.85%. But Exxon fell only 0.24% and Baker Hughes rose 0.55% — the integrated with the least war-levered earnings and the service name with the longest cycle both held. Trump added a policy overhang, saying of Exxon and Chevron: “They’re making too much money based on a shortage. I don’t like it.” The distillate crack gave back with heating oil −5.56% and gasoline −4.77%. Expression: if you owned the crack, this is the day it stopped being free — take the convexity off and keep only the physical-tightness leg (heating oil is still +17.27% on the month and +82.32% YTD).

10. The Korean won strengthened on the day the Kospi fell 5%, which retro-validates Friday’s contrarian read exactly. Kospi −5.12% to 6,257.45 after Friday’s record +17.9%, with SK Hynix −6.29% and Samsung Electronics −7% — and USD/KRW fell 0.95% to 1,429.24, i.e. the won strengthened as the index collapsed. Friday the won weakened 1.35% as the index posted its best day in history. A currency that moves inversely to its own equity market on both legs of a round trip is not driven by foreign portfolio flow; it is driven by domestic leverage unwinding and re-levering, exactly as this report argued on Friday. Bloomberg’s Monday follow-up — “Korean Firms Expanded Short-Term Debt Funding Before Market Rout” — names the funding channel. Keep fading Korea beta against Taiwan; the invalidation is unchanged at USD/KRW below 1,420.

Ten green out of eleven, and the one red group is the one that had to be red — this is the widest sector table since the drawdown began, and it is the mirror image of Friday’s. On Friday six groups fell on a +0.70% index day and Technology itself was down 0.49%; Monday Technology rose 1.58% and the leadership was Communication Services +3.82% (Alphabet +4.88%, Meta +6.01%) and Industrials +2.45% (Boeing +8.01%, Eaton +5.55%, Ingersoll Rand +5.28%, Caterpillar +1.97%, Honeywell +1.54%) — note that Finviz classifies Amazon in Consumer Cyclical, not Technology, which is why Consumer Cyclical +2.35% carries the $3tn story rather than the Technology line. Energy −1.26% on a 5.1% crude settlement drop is a beta of roughly 0.25 to the barrel, which is what you would expect when the group is already +31.40% YTD and 12.95 points clear of Technology — the equity market never fully capitalised the war premium, so it does not have to give it all back. Healthcare −0.34% is the AstraZeneca/Bristol fade plus Eli Lilly −2.39% and Merck −1.87%, the second-worst group on the day the sector produced the largest M&A headline of the year. Consumer Defensive −0.41% with Coca-Cola −0.83%, McDonald’s −2.00% and Walmart −0.44% is the standard risk-on funding trade; Utilities +0.05% barely participated despite a 5 bp rally in the 10-year, because Vistra +5.23% and Constellation Energy are AI-power names that trade with the Nasdaq while the regulated names went nowhere. On the week, the composition is now emphatically pro-cyclical: Consumer Cyclical +8.22% and Communication Services +7.19% lead, Utilities −2.64% and Real Estate −1.47% lag.

Reconciliation. All eleven groups reconcile against Friday’s published YTD compounded by Monday’s 1-day move: Energy 1.3310 × 0.9874 = +31.42% vs. 31.40%; Technology 1.1661 × 1.0158 = +18.45% vs. 18.45%; Industrials 1.1054 × 1.0245 = +13.25% vs. 13.20%; Real Estate 1.1125 × 1.0022 = +11.49% vs. 11.51%; Basic Materials 1.0709 × 1.0113 = +8.30% vs. 8.40%; Financial 1.0723 × 1.0086 = +8.15% vs. 8.13%; Consumer Defensive 1.0833 × 0.9959 = +7.89% vs. 7.84%; Healthcare 1.0492 × 0.9966 = +4.56% vs. 4.58%; Utilities 1.0361 × 1.0005 = +3.66% vs. 3.66%; Communication Services 0.9909 × 1.0382 = +2.88% vs. 2.93%; Consumer Cyclical 0.9667 × 1.0235 = −1.06% vs. −1.06%. Maximum deviation 0.10 pt (Basic Materials); median 0.03 pt. Communication Services, which drifted 0.37 pt on Friday, reconciles to 0.05 pt today. Finviz buckets are not official GICS/S&P sector indices.

Upside

• CoreWeave (CRWV) +19.47% to $85.74, range $69.89–$86.20 on 39.2m shares — Truist upgraded to Buy after a 42% pullback, citing AI compute demand and the valuation gap to other neocloud names; Oppenheimer reiterated Outperform with a $150 target (75% upside from the close), expecting Q2 revenue at the high end of guidance with AI demand running “about four times current supply.” The stock traded down 2.6% intraday before the reversal.

• Oracle (ORCL) +9.23% to $141.86, range $130.21–$143.07 on 44.7m shares — the deepened Google Cloud partnership announced 30 July, bringing Gemini models natively into Fusion Applications, AI Agent Studio and NetSuite so Oracle’s installed base can deploy autonomous agents without separately procuring a foundation model.

• Boeing (BA) +8.01% to $233.45 — FAA certification of the 737 MAX 7 after nearly a decade.

• Rocket Lab (RKLB) +8.44% to $70.43 and IREN +8.02% to $39.75 — the high-beta space and AI-datacentre complex, both bought off the morning lows.

• Norwegian Cruise (NCLH) +6.64% to $19.76, United Airlines (UAL) +5.82% to $128.39, Carnival (CCL) +3.33% to $28.74 — the pure fuel-cost trade and the highest-torque long expression of a 5% crude drop.

• IONQ +6.61% to $38.85 and SanDisk (SNDK) +6.07% to $1,288.58 (range $1,121.27–$1,316.45) — SanDisk recovered its entire Friday loss and more, two days before its 8/5 print.

• Corning (GLW) +6.07% to $146.64 — Truist to Buy from Hold, target $175 (reduced from $205, so a valuation-driven upgrade rather than an estimate-driven one).

• Meta Platforms (META) +6.01% to $590.18, Microsoft (MSFT) +4.94% to $487.66, Alphabet A (GOOGL) +4.88% to $373.51 / Alphabet C +4.44%, Amazon (AMZN) +4.59% to $284.04 — the megacap AI complex moving together for the first time in three weeks. Amazon’s $3tn market capitalisation is the milestone.

• SpaceX (SPCX) +5.68% to $114.53 (not an S&P 500 constituent) — ahead of Tuesday’s first quarterly report as a public company and Thursday’s insider lock-up expiry. Short interest is 32.2% of the tradable float (S3), and Bloomberg flags as much as $116bn of stock becoming eligible for sale next month. Sell-side targets: Deutsche Bank buy $255 (135% upside from Friday’s close); Bernstein outperform $239; RBC outperform $225 (“the lock-up expiration is a material overhang” — Ken Herbert); Susquehanna neutral $170.

• Eaton (ETN) +5.55% to $438.23 — Evercore ISI to Outperform from In Line, target $502 from $453 (14.6% upside). Ingersoll Rand (IR) +5.28% to $87.78 — Stifel to Buy from Hold, target $101 from $84 (15.1% upside).

• Vistra (VST) +5.23% to $155.94 and Palo Alto Networks (PANW) +4.61% to $347.14.

• Alibaba ADR (BABA) +4.11% to $127.28 (not an S&P 500 constituent) — Qwen3.8-Max, the company’s largest AI model, with open weights due next week. DeepSeek launched V4 Flash in beta on Friday.

• Sherwin-Williams (SHW) +3.92%, Tesla (TSLA) +3.50% to $322.10, Marvell (MRVL) +3.34%, Clorox (CLX) +2.86% (reported after the close), Tyson Foods (TSN) +2.85% to $59.61 (range $54.53–$59.91) — Tyson missed at $0.99 adjusted EPS versus $1.01 consensus on $13.868bn revenue versus $14.07bn, but rallied on the USDA resuming Mexican cattle imports and management’s view that momentum continues “across all of its businesses, even beef.”

• NVIDIA (NVDA) +2.96% to $206.70 on 105.9m shares, American Express +2.52%, Home Depot +2.43%, Netflix +2.25%, Nike +2.23%, Palantir (PLTR) +2.10% to $125.65 ahead of its print, Walt Disney +2.03%, Caterpillar +1.97% into Tuesday’s 6:30 a.m. report, AMD +1.75% to $484.46 into Tuesday’s 4:15 p.m. report, Honeywell +1.54%, IBM +1.12%, Salesforce +1.05%, Intel +0.89%, Goldman Sachs +0.85%, Micron +0.79% to $829.50, Broadcom +0.76%.

• Imax hit an all-time high (not an S&P 500 constituent) — CEO Rich Gelfond told CNBC that Christopher Nolan’s “The Odyssey” is tracking at almost double “Oppenheimer.” AMC +6% on a $355m opening weekend for “Spider-Man: Brand New Day.”

Downside

• Marriott (MAR) −7.05% to $346.56, range $344.13–$365.47 — revenue $7.07bn versus a $7.17–$7.26bn consensus range; adjusted EPS beat, but Q3 guidance of $2.74–$2.82 came in below the $2.87 estimate and international RevPAR fell 0.5% with the Middle East down 43%. Full-year RevPAR guidance was raised; the market ignored it. The most direct read that the war’s demand cost is still being discovered even as its price premium unwinds.

• eBay (EBAY) −6.03% to $107.13, range $107.08–$112.19 — Wells Fargo downgrade to Underweight from Equal Weight; closed on the low.

• Marathon Petroleum −2.98%, Occidental −2.80%, NXP Semiconductors −2.18% (UBS to Neutral from Buy, target $270 from $305), Diamondback −2.07% into its own after-close print, EOG Resources −2.02%, Chevron −1.85%, Valero −1.71%, ONEOK −1.67%, Williams −1.51%, Devon −1.21%, Halliburton −1.12%, ConocoPhillips −1.10%, Exxon Mobil −0.24%.

• Eli Lilly (LLY) −2.39% to $1,121.36, range $1,109.15–$1,161.94 — closed near the low ahead of Wednesday’s 6:45 a.m. report. Merck −1.87% into Tuesday’s print. Amgen −1.63% into Tuesday after the close. Vertex −1.34% into its own after-close print.

• McDonald’s −2.00% (range $265.23–$276.33, closed at the low) into Tuesday’s 7:00 a.m. report; Starbucks −1.79% despite a Melius upgrade to Hold from Sell; Best Buy −1.17%; Coca-Cola −0.83%; Johnson & Johnson −0.76%; Walmart −0.44%.

• Apple (AAPL) −1.78% to $303.42 on 71.9m shares, range $302.56–$311.80 — the only Magnificent Seven name red, down 0.8% at midday and weaker into the close. Still +13% in 2026 after last week’s 7.35% slide on Tim Cook’s advanced-chip and memory supply warning.

• ON Semiconductor −1.48% into its own after-close print; SBA Communications −0.80% likewise; Travelers −0.15%, Visa −0.13%, Cisco −0.12%.

• GameStop −12.25% and Sportradar −15.13% (neither an S&P 500 constituent) were the largest liquid decliners on the tape.

Notable corporate and macro items

• Palantir (PLTR) after the close: revenue $1.94bn, +92.8% y/y, versus a $1.81bn StreetAccount estimate; EPS $0.41 versus $0.33; U.S. commercial revenue $764m, +149% y/y; GAAP operating income $912m at a 47% margin; adjusted operating income $1.19bn at 62%; Rule of 40 at 155%. Full-year guidance raised to $8.150–$8.158bn (+82% y/y), U.S. commercial to above $3.424bn (+134%), adjusted FCF $4.5–$4.7bn. Bloomberg noted the shares are still “down some 40% since hitting a peak in November.” The price reaction belongs to Tuesday’s session and is not in this report’s closing data.

• M&A: KKR agreed to buy Integer Holdings in an all-cash deal valued at about $5.7bn; Williams agreed to buy Momentum Midstream for $5.5bn; AstraZeneca/Bristol Myers Squibb talks reported by the FT at roughly $400bn combined. ArcelorMittal +2% on a deepened Microsoft Azure partnership.

• Earnings-season scoreboard: BofA Securities reports the Q2 EPS beat rate at 77%, the highest going back to 2021, with S&P 500 Q2 earnings tracking +27% y/y excluding Alphabet and Amazon mark-ups and +45% including them. Bloomberg’s count through Friday’s close: of 307 S&P 500 companies reported, 86% beat on EPS; 68% beat on sales, 15% missed. Jefferies’ Jeffrey Favuzza: “We count 95 tech earnings this week across the US with more than $1 billion in market cap.”

• Strategy (MSTR) sold 1,638 BTC for $104.7m between 27 July and 2 August at an average $63,957, using roughly half to fund preferred dividends and the rest to repurchase STRC; it raised $290.6m via ATM. U.S. spot bitcoin ETFs saw $265.37m of net outflows on Friday, taking the week to −$61.5m.

• Valuation commentary: Oppenheimer’s John Stoltzfus noted the S&P 500 ended last week at 19.7× forward earnings, 2.5% below its five-year average: “Stocks are getting relatively cheaper.” RBC’s Lori Calvasina said valuations on the Nasdaq 100, S&P 500 and the technology sector “are starting to look reasonable again,” while warning that “the path for stocks will not be a linear one.” SocGen’s Alain Bokobza took the other side: “A second wave of US tariffs, the accelerating AI and infrastructure capex cycle, renewed oil-price volatility and persistently large fiscal deficits across developed economies all point to a more inflationary backdrop than markets currently discount.” Citadel Securities’ Shah said markets are “questioning Warsh’s plans” while the firm’s public line was that bull-market drivers are “firmly intact”; Wolfe Research’s Stephanie Roth wrote that Warsh “faces lingering questions about the central bank’s credibility” and that “the data may yet bail him out,” with Jackson Hole offering “a reset.”

• William Blair added GE Vernova to its conviction list with an Outperform rating: “the market still underestimates the level by which electrical infrastructure and the U.S. power grid need to be built out… low cost models like Kimi K3 lower the ROI hurdle for enterprise AI deployment.” Shares +1%.

Times are ET. Every day page from Tuesday 8/4 through Friday 8/14 was re-pulled from Earnings Whispers this session (/1 = before open, /2 = after close). Non-S&P-500 names are excluded, as are names whose index membership could not be verified conservatively. Re-verify times and membership against company IR before trading any date.

BMO: Loews (L) 6:00 — closed +0.21%; Marriott (MAR) 7:00 — closed −7.05% at $346.56 on a $7.07bn revenue miss and a soft Q3 guide; Tyson Foods (TSN) 7:30 — closed +2.85% despite a $0.99 vs. $1.01 EPS miss.

AMC: SBA Communications (SBAC) 4:00 — −0.80% into the print; Vertex Pharmaceuticals (VRTX) 4:00 — −1.34%; Diamondback Energy (FANG) 4:00 — −2.07%; Palantir (PLTR) 4:05 — revenue $1.94bn, +92.8% y/y vs. $1.81bn; EPS $0.41 vs. $0.33; FY26 guide raised to $8.150–$8.158bn; ON Semiconductor (ON) 4:05 — −1.48%; Alexandria Real Estate (ARE) 4:10; Clorox (CLX) 4:10 — +2.86%; ONEOK (OKE) 4:15 — −1.67%; Williams (WMB) 4:15 — −1.51%; also announced a $5.5bn acquisition of Momentum Midstream.

BMO: Archer-Daniels-Midland (ADM) 6:00, Ball (BALL) 6:00, DuPont (DD) 6:00, Gartner (IT) 6:00, Henry Schein (HSIC) 6:00, Leidos (LDOS) 6:00, Revvity (RVTY) 6:00, Waters (WAT) 6:00, Apollo Global (APO) 6:30, Caterpillar (CAT) 6:30, IDEXX (IDXX) 6:30, Kimberly-Clark (KMB) 6:30, Merck (MRK) 6:30, Zebra (ZBRA) 6:30, Aptiv (APTV) 6:45, Marathon Petroleum (MPC) 6:45, Pfizer (PFE) 6:45, Kimco (KIM) 6:50, AMETEK (AME) 6:55, Broadridge (BR) 7:00, Duke Energy (DUK) 7:00, McDonald’s (MCD) 7:00, NRG Energy (NRG) 7:00, Rockwell Automation (ROK) 7:00, TransDigm (TDG) 7:15, Cummins (CMI) 7:30, FIS (FIS) 7:30, PSEG (PEG) 7:30, Sysco (SYY) 8:00, W.W. Grainger (GWW) 8:00, Progressive (PGR) 8:15, Expeditors (EXPD) 8:30, Pinnacle West (PNW) 8:35. CAT at 6:30 is the most important industrial print of the week; MPC at 6:45 is now the distillate-crack referendum after Monday’s −2.98%.

AMC: Amgen (AMGN) 4:00, Booking Holdings (BKNG) 4:00, Gilead (GILD) 4:00, Wynn Resorts (WYNN) 4:00, Arista Networks (ANET) 4:05, DaVita (DVA) 4:05, Devon Energy (DVN) 4:05, Emerson Electric (EMR) 4:05, Jacobs (J) 4:10, Match Group (MTCH) 4:10, AMD (AMD) 4:15 — the single most important print of the week for the semiconductor complex, Celanese (CE) 4:15, Healthpeak (DOC) 4:15, IFF (IFF) 4:15, Mosaic (MOS) 4:15, Prudential Financial (PRU) 4:20, Assurant (AIZ) — “After Close,” no specific time. Also on this page: SpaceX (SPCX) “After Close” — membership not verified conservatively and therefore excluded from the S&P list, but the most-watched non-index print of the week.

BMO: EOG Resources (EOG) 1:25 AM, Cencora (COR) 6:30, CVS Health (CVS) 6:30, NiSource (NI) 6:30, Zimmer Biomet (ZBH) 6:30, Iron Mountain (IRM) 6:45, Eli Lilly (LLY) 6:45, Global Payments (GPN) 6:55, Uber (UBER) 6:55, CDW (CDW) 7:00, Charles River Labs (CRL) 7:00, Insulet (PODD) 7:00, Kraft Heinz (KHC) 7:00, Phillips 66 (PSX) 7:00, Walt Disney (DIS) — “Before Open,” no specific time. Owens Corning (OC), BorgWarner (BWA) and United Therapeutics (UTHR) were on the prior verified pull for this date but did not appear in this session’s listing — confirm with company IR.

AMC: Axon (AXON) 4:00, Expedia (EXPE) 4:00, MercadoLibre (MELI) 4:00, Western Digital (WDC) 4:00, AppLovin (APP) 4:05, Block (XYZ) 4:05, Corpay (CPAY) 4:05, DoorDash (DASH) 4:05, eBay (EBAY) 4:05, Paycom (PAYC) 4:05, Realty Income (O) 4:05, SanDisk (SNDK) 4:05, Solventum (SOLV) 4:05, McKesson (MCK) 4:10, Motorola Solutions (MSI) 4:10, Albemarle (ALB) 4:15, MetLife (MET) 4:15, News Corp B (NWS) 4:15, Occidental (OXY) 4:15, Texas Pacific Land (TPL) 4:15, News Corp A (NWSA) 4:25, CF Industries (CF) 4:30, Host Hotels (HST) 4:30, Steris (STE) 4:30, Atmos Energy (ATO) 4:35, Allstate (ALL) 5:10. WDC and SNDK together remain the memory-pricing referendum that adjudicates the §12 dispersion trade.

BMO: EPAM (EPAM) 6:00, Targa Resources (TRGP) 6:00, Becton Dickinson (BDX) 6:30, Kenvue (KVUE) 6:30, Molson Coors (TAP) 6:30, Teleflex (TFX) 6:30, Viatris (VTRS) 6:55, ConocoPhillips (COP) 7:00, Evergy (EVRG) 7:00, Howmet (HWM) 7:00, Keurig Dr Pepper (KDP) 7:00, Warner Bros. Discovery (WBD) 7:00, Zoetis (ZTS) 7:00, Constellation Energy (CEG) 7:05, Parker-Hannifin (PH) 7:30, Sempra (SRE) 7:55, APA (APA) 8:00, Fox Class B (FOX) 8:00, Ralph Lauren (RL) 8:00, Fox Corporation (FOXA) — “Before Open.” Datadog (DDOG), Fiserv (FI) and Cheniere (LNG) were on the prior verified pull for this date but did not appear this session — confirm with company IR.

AMC: Airbnb (ABNB) 4:00, Akamai (AKAM) 4:00, Texas Roadhouse (TXRH) 4:00, Aflac (AFL) 4:05, Gen Digital (GEN) 4:05, ResMed (RMD) 4:05, Monster Beverage (MNST) 4:10, Republic Services (RSG) 4:10, AIG (AIG) 4:15, Microchip Technology (MCHP) 4:15, Consolidated Edison (ED) 4:30. The Trade Desk (TTD), DraftKings (DKNG) and Reinsurance Group (RGA) were on the prior verified pull and did not appear this session — confirm with IR.

BMO: Vistra (VST) — “Before Open”; Take-Two Interactive (TTWO) 7:00; PPL (PPL) 7:30. MarketAxess (MKTX), carried on the prior pull, is absent for a second consecutive session — treat as removed and confirm with IR (MKTX is an announced acquisition target of ICE).

AMC: the reviewed after-close page for 8/7 lists only non-S&P-500 names (Hawaiian Electric, Galaxy Gaming) — no S&P 500 after-close reporters are published for that date, unchanged for a third consecutive edition.

Mon 8/10. BMO: the reviewed before-open page lists no S&P 500 reporters. AMC: Simon Property Group (SPG) 4:05.

Tue 8/11. BMO: Cardinal Health (CAH) 6:45. AMC: Super Micro Computer (SMCI) 4:05.

Wed 8/12 — CPI day (8:30). BMO: Amcor (AMCR) 6:00. AMC: Cisco (CSCO) 4:05.

Thu 8/13 — PPI day (8:30). BMO: Tapestry (TPR) 6:45. AMC: Applied Materials (AMAT) 4:00.

Fri 8/14 — retail sales and Michigan preliminary. Neither the before-open nor the after-close page lists any S&P 500 reporter.

Changes vs. the prior calendar (7/31 report). Confirmed removal: MarketAxess (MKTX) from Fri 8/7 BMO, absent for a second consecutive session. Additions to Wed 8/5 AMC: Motorola Solutions (MSI) 4:10, News Corp A/B (NWSA 4:25 / NWS 4:15), McKesson (MCK) 4:10, MetLife (MET) 4:15, Realty Income (O) 4:05, Steris (STE) 4:30, Solventum (SOLV) 4:05 — this after-close block is now the densest of the week at 26 S&P 500 names. Additions to Thu 8/6 BMO: Warner Bros. Discovery (WBD) 7:00, Zoetis (ZTS) 7:00, Sempra (SRE) 7:55, Ralph Lauren (RL) 8:00, Teleflex (TFX) 6:30, Parker-Hannifin (PH) 7:30. Names on the prior verified pull that did not appear this session and are flagged rather than dropped: OC, BWA, UTHR (8/5 BMO); DDOG, FI, LNG (8/6 BMO); TTD, DKNG, RGA (8/6 AMC) — Earnings Whispers renders a bounded top-of-page listing, so absence from one capture is not proof of removal. The forward calendar collapses next week: across the ten Aug 10–14 day pages there are only six S&P 500 reporters versus roughly 150 this week, and Cisco (8/12 AMC) and Applied Materials (8/13 AMC) are the only two that move an index — both landing on inflation-print days.

Curve spreads. 2s10s +45 bp (−2 bp d/d, +11 bp w/w). 3M10Y +79 bp (−13 bp d/d, +10 bp w/w). 2s30s +98 bp (−1 bp d/d, +17 bp w/w).

Name the shape: a belly-led bull-flattening from the 2-year out, bolted onto an 8 bp bear move in the 3-month bill — two unrelated events on one curve, and only one of them is about the Fed. From the 2-year outward the entire curve rallied, with the maximum move in the 5s/7s/10s/20s at −5 bp against the 2Y at −3 bp and the 30Y at −4 bp. That is Friday’s belly leadership running in reverse: the belly is where the market prices the expected path of policy plus the inflation compensation attached to it, and a 5.1% crude settlement drop takes both out at once. It is not a term-premium event — the 30-year fell less than the 10-year, so the compensation for holding duration barely moved. It is not a growth scare either: ISM Manufacturing printed 55.6, the strongest since May 2022, and the 2-year only fell 3 bp. This is a clean, single-factor oil-and-inflation-compensation rally, and its honest size is five basis points.

The 3-month is the exception and it is the most interesting number on the page. 3M +8 bp to 3.91% while the 10-year fell 5 bp — a 13 bp flattening of 3M10Y in one session, with 2M +2 bp, 4M +2 bp and 6M +4 bp confirming the cheapening is concentrated in the belly of the bill curve rather than at the overnight point (1M only +1 bp). This is a financing signal, not a policy signal: the 3-month bill now yields 25 bp over overnight SOFR (3.91% vs. 3.66%), an unusually wide bill–OIS box, alongside a record $3.205tn SOFR volume print, a SOFR fix 1 bp above IORB, a 99th percentile that touched the 3.75% top of the target range, and reserve balances down $158bn from their 15 July peak (§9). Bills cheapening into the refunding window while reserves drain is a textbook supply-and-plumbing combination, and it says nothing about September.

The week is the more important frame, and the week is still a bear-steepener. Against the same weekday one week earlier: the front end is 6 to 11 bp lower (4M −11, 2M −8, 6M −8, 1Y −7, 2Y −6) while the long end is 5 to 11 bp higher (10Y +5, 20Y +8, 30Y +11), with the 5-year exactly unchanged as the pivot. 2s30s has steepened 17 bp in a week and 2s10s 11 bp. Monday clawed back roughly a third of Friday’s long-end damage and none of the week’s. The structural trade — the market pricing out tightening at the front while demanding more compensation to own the back — is intact, and one dovish oil headline did not break it.

Vendor reconciliation. Bloomberg’s 4 p.m. real-time mark was 4.68% against the official par close of 4.70%. The 2 bp gap is the usual baseline/timing artefact — Treasury’s par curve is a 3:30 p.m. bid-side construction fitted across the coupon universe; Bloomberg’s board is a single on-the-run print at 4 p.m. Both agree the 10-year fell about 5 bp; there is no level dispute. All d/d and w/w changes above are computed on the official series so the basis is internally consistent.

Current week (Aug 3–7) — jobs week: released

Take on ISM. This is the rare print that is good for both legs of a 60/40 book: growth accelerating (55.6, employment back in expansion) while the price index falls 1.9 points. It is also why the front end barely rallied — a 52.8 employment index two days before payrolls is not a print that lets the market add rate cuts. The level of prices paid at 71.1 is the caveat and it is explicitly oil-linked, which makes Monday’s 5% crude drop a genuine forward-looking disinflation input rather than a one-day artefact.

Remaining this week (NY Fed August indicator calendar, re-verified; all times ET)

Next week (Aug 10–14) — the inflation week

Current target range: 3.50–3.75% (held 9–3 on 7/29; Hammack, Kashkari and Logan dissented for +25 bp; IORB 3.65%). Chair: Kevin Warsh. Next FOMC: Wednesday, September 16, 2026, 2:00 PM ET — 43 days away. CME September card read live and stamped 3 Aug 2026 04:58:37 CT (contract ZQU6, expiry 30 Sep 2026, mid price 96.2850, prior volume 33,079, prior open interest 221,602). Investing.com cards stamped Aug 03, 2026 05:45 PM EDT.

CME FedWatch headline — September 16, 2026 meeting

Column sums: 100.0 / 100.0 / 100.1 / 100.0 — the 1-WEEK column rounds to 100.1.

Provenance of every column, stated plainly — and it improved this session. Unlike the last several editions, CME’s tool published a complete numeric four-column table today, dated 1 DAY = 31 Jul 2026, 1 WEEK = 27 Jul 2026, 1 MONTH = 2 Jul 2026, so all four columns above are read directly off CME and none is carried, estimated or reconstructed from a prior edition. That exposes a real discrepancy worth flagging rather than hiding: CME’s own 1-DAY snapshot for 31 July reads 33.0% hold / 67.0% hike, whereas this report published 27.9% / 72.1% on Friday from a live read of the same tool at roughly 6:05 p.m. ET. The reconciliation is timing, not error — CME’s historical columns are end-of-day settlement snapshots, while Friday’s live read was taken after the fed funds futures settlement window on a month-end session and captured an intraday state that did not survive to the official close. The rule that follows: a live CME read taken after 5:00 p.m. ET should be treated as indicative; the number that persists into the historical series is the settlement figure. Friday’s figure is corrected here to CME’s own 33.0/67.0.

Reconciling CME against Investing.com — the gap has collapsed from 6.2 points to 1.1, and the reason is instructive. For the same September meeting, CME puts the hold at 32.8% and the +25 bp at 67.2%; Investing.com puts them at 33.9% and 66.1% — a 1.1-point gap. On Friday the same comparison was 6.2 points, traced to a 2.0 bp difference in the quoted ZQU6 contract price (CME 96.2700 vs. Investing.com 96.290). Today the two vendors quote the identical contract: CME’s mid price is 96.2850 and Investing.com’s future price is 96.285. With the price input identical, the entire residual gap is methodology: CME day-weights the September contract (the meeting falls on the 16th, so only about 14 of the month’s 30 days carry the post-meeting rate) and anchors on the realised effective rate (EFFR is fixing at 3.63%, half a basis point above the 3.625% range midpoint), while Investing.com maps the contract price on a simpler basis. Calibration from the two observations: roughly 3 points of headline probability per basis point of ZQU6 price difference, with about 1 point of irreducible methodology spread when the prices agree. Use one vendor’s columns consistently within any comparison; do not mix them.

(1) 2026 meeting distributions (Investing.com current / [prev-day] / [prev-week])

Row sums: September 100.0 / 100.0 / 100.0; October 99.9 / 100.1 / 100.1; December 100.0 / 100.1 / 100.0. No easing is priced at any 2026 meeting in any column — 0.0% below 3.50% throughout, for the eleventh consecutive session. Contract prices: Sep 96.285, Oct 96.190, Dec 96.045.

Multi-day momentum — September stood still while everything behind it eased. On CME’s cumulative basis the September hike arc reads 54.1% a month ago → 81.5% a week ago → 67.0% a day ago → 67.2% now: unchanged on the day (+0.2 pt), down 14.3 points on the week, up 13.1 points on the month. That is the striking feature of the session: a 5% crude collapse and a 1.9-point drop in ISM prices paid moved the September card by two-tenths of a point. The easing showed up further out. October cumulative fell from 76.9% to 76.2% with the +50 bp bucket down from 21.2% to 19.7%; December-2026 cumulative fell from 87.5% to 86.4%, with ≥+50 bp down from 46.7% to 44.0% (−2.7 pt) and ≥+75 bp from 9.7% to 8.5%; the December hold rose from 12.6% to 13.6%. Week-over-week: December-2026 ≥+50 bp is 44.0% against 59.1%, and ≥+75 bp 8.5% against 20.9%. The distribution keeps compressing toward exactly one hike, and Monday took another slice off the second one without touching the first.

(2) 2027 meeting path (modal range, probability, cumulative above/below current 3.50–3.75%)

Row sums: Jan 99.9; Mar 100.0; Apr 99.9; Jun 100.0; Jul 99.9; Sep 100.0; Oct 100.0; Dec 100.0. Contract prices: Jan 96.010, Mar 95.940, Apr 95.900, Jun 95.865, Jul 95.865, Sep 95.885, Oct 95.900, Dec 95.940.

Two changes worth naming. First, January 2027’s mode flipped back down a step, from 4.00–4.25% on Friday to 3.75–4.00% today — and it flipped by one-tenth of a point (36.9% vs. 36.8%). That is a coin-toss, not a conviction shift, but it is the second time in three sessions the January card has changed its mind, and it is the meeting where the market decides whether the Fed does one hike or two. Second, the terminal rate implied by the forward strip fell about 4 bp. Contract prices bottom at 95.865 in June and July 2027 and then rise to 95.940 by December 2027, drawing a peak policy rate near 4.14% in mid-2027 — against 95.825 and roughly 4.18% on Friday. Cumulative above-range probability peaks at 93.2% at the June 2027 meeting (94.6% on Friday) before decaying to 83.2% by December 2027 (86.4% on Friday). The first meaningful easing probability still appears late and small: 0.2% below 3.50% by July 2027, 0.9% by September, 1.6% by October, 3.3% by December — every one larger than Friday’s reading, the cleanest single measure of how much the oil move mattered to the back end.

(3) Year-end 2026 probability ladder (December 9, 2026) — current / [prev-day] / [prev-week]

Cumulative: ≥+25 bp 86.4% / [87.5%] / [91.7%]; ≥+50 bp 44.0% / [46.7%] / [59.1%]; ≥+75 bp 8.5% / [9.7%] / [20.9%]; ≥+100 bp 0.0% / [0.0%] / [3.0%]; any cut 0.0% in all three columns. Column sums 100.0 / 100.1 / 100.0.

Year-end 2027 probability ladder (December 8, 2027) — current / [prev-day] / [prev-week]

Cumulative: ≥+25 bp 83.2% / [86.4%] / [84.8%]; ≥+50 bp 54.2% / [60.5%] / [58.9%]; ≥+75 bp 23.4% / [29.2%] / [28.8%]; ≥+100 bp 6.1% / [8.8%] / [9.1%]; any cut 3.3% / [2.5%] / [3.3%]. Column sums 100.0 / 99.9 / 100.0.

(4) Rate-path interpretation

(a) Investment-grade and high-yield spreads

As-of date, stated up front: the ICE BofA option-adjusted spread series on FRED are published with a lag and the latest observation available at the time of writing is 30 July 2026, not 3 August. FRED’s series page for BAMLC0A0CM shows “2026-07-30: 0.80, Updated: Jul 31, 2026 8:57 AM CDT, Next Release Date: Aug 3, 2026” — the 31 July observation had not posted to the data table by the U.S. evening. Everything in this table is a T−2 business-day read; same-day direction is inferred separately and stated below.

Read the CCC line, because it is the only one doing anything. IG at 80 bp is one basis point wider than where it started the year and one basis point wider than a week ago. That is not a credit market with a view. High yield at 284 bp is 3 bp wider YTD. But CCC & lower at 1,006 bp is 121 bp wider on the year and 15 bp wider on the week, and it broke 1,000 for the first time in this cycle on 27 July. The entire deterioration in U.S. corporate credit in 2026 has occurred in the bottom rating bucket, while the index-level spread has not moved at all. The CCC/IG ratio is now 12.6× against 11.2× at the start of the year — the signature of an idiosyncratic default cycle running underneath a healthy index, with refinancing stress concentrated in the most levered, most floating-rate borrowers: precisely the cohort a 3.50–3.75% policy rate with a 67% chance of going higher damages most. All three series tightened on 30 July, and the same-day direction on 3 August was almost certainly tighter still — a 1.5% equity rally with the Russell outperforming and crude down 5% is not a session in which credit widens — but that inference is not published as a number.

(b) Money-market and funding plumbing

Publication basis: the New York Fed publishes reference rates at approximately 8:00 a.m. ET for the prior business day, so the most recent published fixings are for Friday 31 July. ON RRP take-up is available through Monday 3 August; reserve balances are weekly through the week ending 29 July.

The plumbing is the quiet story of this report and it deserves the same attention as the Fed section. Four facts line up: SOFR fixing above IORB, a 99th percentile at the top of the target range, record volume, and reserves down $158bn in a fortnight with ON RRP already at zero. Individually each is a month-end artefact. Together they describe a system in which the reverse repo facility has stopped providing the marginal collateral cushion and reserves are being drained faster than bills are being absorbed — which is exactly what an 8 bp one-day cheapening of the 3-month bill against a 5 bp rally in the 10-year looks like from the other side. This is not yet a stress signal — a 1 bp SOFR–IORB spread is nothing like September 2019 — but it is the first time this year that all four have printed together, and the next quarter-end is 30 September, two weeks after the FOMC. Watch the SOFR–IORB spread through mid-August: if it holds above IORB on non-month-end days, the Fed’s balance-sheet trajectory becomes a live variable in the September meeting in a way it is not currently priced to be.

(c) Rates volatility and swap spreads

This is the divergence to trade and it is the natural companion to §6 and §8. Rates volatility has risen 17% in eleven sessions while equity volatility has fallen 23% in three. The bond market is pricing a live policy and term-premium problem — a 67% September hike, a 30-year that made a 19-year closing high three sessions ago, a curve that has bear-steepened 17 bp in 2s30s in a week — while the equity market has taken the VIX to 15.86, below where it sat before the July drawdown began. Goldman’s Lee Coppersmith said it plainly on Monday: “Treasury volatility has started to re-accelerate, real yields remain near cycle highs, and historically that’s not an environment where equity volatility stays structurally compressed.”

(d) Issuance, leveraged loans and private credit

• Primary supply and the forward calendar. No aggregate IG/HY daily volume figure was published on any source rendered this session; what is observable is the acquisition financing pipeline Monday created. KKR agreed to buy Integer Holdings in an all-cash deal valued at about $5.7bn, and Williams agreed to buy Momentum Midstream for $5.5bn — two announcements in one day implying roughly $11bn of new funding need in the second half, and Williams is an investment-grade midstream issuer that has historically termed out acquisition bridges in the public IG market. The AstraZeneca/Bristol Myers combination, if it happened, would be an order of magnitude larger. Against an IG index at 80 bp, that supply is being announced into the tightest credit market since the start of the year.

• Leveraged loans. The Morningstar LSTA US Leveraged Loan index level was not obtainable from a primary source this session. The observable read-across is the CCC OAS at 1,006 bp, +121 bp YTD, which is where the loan market’s floating-rate stress surfaces in the bond market: loan borrowers are disproportionately single-B and CCC, and a policy rate with a 67% chance of rising and 0% chance of falling in 2026 keeps their interest burden climbing regardless of what the index-level spread does.

• Bank CDS and financial credit. No quoted level was obtainable. The equity read was benign — Financials +0.86%, Goldman Sachs +0.85%, JPMorgan +0.24% — and UBS agreed to pay $125m over anti-money-laundering violations (Bloomberg), a headline-risk item rather than a credit event.

• Named private-credit and leverage watch items — three landed on Monday and they belong together. (i) “Korean Firms Expanded Short-Term Debt Funding Before Market Rout” (Bloomberg) — the funding channel behind the Kospi’s round trip, and the most concrete evidence yet that the Korean equity move was levered rather than allocated. (ii) WSJ: “His Wedding Guests Were Arriving—Just as His $45 Billion Fund Was Falling Apart,” and “Citadel Buys Situational Awareness’s Stock Portfolio After Big Losses in AI” — a $45bn fund unwinding an AI book into a Citadel bid is the single-manager version of the de-grossing JPMorgan quantified, and the mechanism by which July’s semiconductor liquidation actually happened. Bloomberg’s framing: “Situational Awareness’ 439% Return Was the Warning.” (iii) Bloomberg Odd Lots, “Why Private Credit Got Entangled With Insurance” — annuity balance sheets funding illiquid private loans is the transmission line that turns a CCC default cycle into a regulated-liability problem.

Take — the dollar’s non-response is the story, and the Asian crosses have now flipped twice in two sessions. The DXY rose eight basis points on a day the Dow closed at a record, ISM Manufacturing printed its best number since May 2022, and the oil-driven terms-of-trade shock ran in America’s favour. A currency that cannot rally on that combination is not being traded on growth differentials; it is being traded on the credibility question Wolfe Research and Citadel Securities both named on Monday, and on a 10-year that fell 5 bp. The dollar’s problem is that good U.S. data now lowers U.S. yields (because it arrives with cheaper oil) rather than raising them. That is a materially different regime from the first half of the year, and it is why the DXY is stuck at 100 with a −1.52% weekly print.

The two crosses carrying information are USD/KRW and USD/TWD, and both inverted from Friday. On Friday the won fell 1.35% as the Kospi rose 17.9%; on Monday the won rose 0.95% as the Kospi fell 5.12%. On Friday the Taiwan dollar firmed 0.30% as the Taiex rose 8%; on Monday it weakened 0.31% as TSMC fell 2.06%. Read the pair of pairs together: Taiwan’s currency moves with its equity market and Korea’s moves against it. That is the cleanest possible statement that Taiwan’s semiconductor rally is foreign-inflow-driven and Korea’s was domestic-leverage-driven — and Bloomberg’s Monday story, “Korean Firms Expanded Short-Term Debt Funding Before Market Rout,” supplies the funding evidence. The trade from Friday is unchanged and has now been confirmed on both legs of a round trip: express Asian semiconductor exposure through Taiwan, fade it through Korea. Invalidation remains USD/KRW below 1,420 on a rising Kospi.

The yen is the one live policy trade. Japan’s finance ministry confirmed the Friday intervention was coordinated with the U.S. Treasury — a rare joint operation — with Finance Minister Satsuki Katayama saying Tokyo remains “attentive and in close communication with counterparts at U.S. Treasury.” The sell side is uniformly unconvinced: UBS’s Teck Leng Tan and Dominic Schnider — “Japan’s policy mix remains unlikely to generate sustained yen strength”; HSBC — “Unless we see much faster BoJ rate hikes, and the government taking a clearer stand on the JPY… we still lack confidence in projecting a downtrend for USD-JPY.” WSJ’s What’s News agreed: “Investors Remain Skeptical About the Yen After U.S. Intervention.” The asymmetry for a U.S. investor is that an intervention floor without a policy change is a free option to sell — but the option is expensive, because the second intervention is always larger than the first.

Quote basis. All pairs are TradingEconomics spot marks pulled at the U.S. evening. Most carry an “Aug/03” date stamp; DXY, USD/CAD and USD/SGD carried a “06:23” time stamp, which is the Tuesday Asia session, so those are marginally forward of the New York close. The USD/JPY line is flagged: TradingEconomics shows 157.395, identical to the figure it published for Friday, with a −0.02% daily change, while CNBC reported the yen at 157.70 on Monday. The vendor’s yen quote appears not to have refreshed for the Monday session and should not be relied on for the daily change; the level is corroborated to within 0.2% by CNBC, and the weekly figure is consistent with the intervention narrative.

Basis caveats, stated up front. Crude is quoted on Bloomberg’s official settlement basis — WTI September settled −5.1% at $80.34 and Brent October −4.7% at $83.77 — because those are exchange settlements rather than board marks. The TradingEconomics board printed $79.944 (−5.58%) for WTI and $83.499 (−5.04%) for Brent, which agree with Bloomberg to within 0.5% and 0.3%. That is a material improvement on Friday, when the same vendor’s crude quotes disagreed with Bloomberg by more than $2 and were withheld; today they are internally consistent and are used for the weekly, monthly and YTD columns only, with the daily change taken from the settlement. The column order was verified programmatically as Price | Chg | %Chg | Weekly | Monthly | YTD | YoY | Date before any figure was quoted. Metals, natural gas and refined products are TradingEconomics board marks with an Aug/03 stamp. The gold line carries a genuine vendor gap that is not an error: TradingEconomics spot shows $4,051.75 (+0.22%) while the Bloomberg /markets board shows $4,109.50 (+0.46%) — a $57.75, or 1.43%, difference consistent with the futures-versus-spot carry on a deferred COMEX contract at a 4%+ funding rate. Spot is used in the table; the Bloomberg futures line is disclosed so the reader can reconcile. “Front month” remains genuinely ambiguous across the roll — September WTI and October Brent are the actively quoted contracts.

Each idea states the expression, the catalyst that resolves it, and the invalidation. Claude is not a licensed financial advisor; sizing and suitability are the reader’s own.

1. Trade the semiconductor complex through dispersion, not direction — two full reversals in two sessions is a basket being worked. Expression: long AMD and NVDA against short MU and SNDK, beta-weighted rather than dollar-neutral, and buy the AMD straddle rather than picking a side into Tuesday’s print. Why: Friday the SOX traded +5.15% intraday and closed +0.07%; Monday it traded −3.44% intraday and closed +1.06%. MU printed a $930.88 high Friday and a $770.10 low Monday and closed both sessions within 1% of $826. A complex that round-trips 8% in both directions inside 48 hours has no information in its direction and all of its information in its dispersion. JPMorgan’s Panigirtzoglou says the de-grossing is largely done, which caps the downside; the memory-pricing question is unresolved, which caps the upside. Catalyst: AMD Tue 8/4 4:15 p.m., WDC and SNDK Wed 8/5, Microchip Thu 8/6. Invalidation: an AMD data-centre number that lifts NVDA and MU by similar percentages — correlation going back to one, which kills the pair.

2. Stay in the 5s30s steepener; Monday gave a better entry, not a reason to leave. Expression: 5s30s steepener in cash or futures; alternatively pay 5-year and receive 30-year in swaps. Why: Monday’s rally was belly-led (5Y, 7Y, 10Y, 20Y all −5 bp against 30Y −4 bp and 2Y −3 bp) but the week is still emphatically a bear-steepener (front end −6 to −11 bp, 10Y +5, 20Y +8, 30Y +11, 2s30s +17 bp). Four consecutive weeks of the front end pricing out tightening while the long end cheapens anyway is a term-premium trend, and a single oil headline does not reverse a term-premium trend. The §9 configuration supports it: MOVE +17% in eleven sessions against a VIX at 15.86. Catalyst: ISM Services prices paid Wed 10:00 a.m., payrolls Fri 8/7, CPI Wed 8/12. Invalidation: two consecutive soft inflation prints, which would let the long end retrace and flatten 5s30s back through +70 bp; or a 30-year auction that stops through.

3. Express the disinflation view in December-2026 fed funds, not September. Expression: long ZQZ6 (or receive December-2026 fed funds / SR3Z6) rather than taking a view on the September contract. Why: September is 67.2% priced and moved 0.2 points on a 5% crude drop; December ≥+50 bp is 44.0% and moved 2.7 points on the same news, and 15.1 points on the week. The September hike is close to a decided question; the second hike is the live one and it carries all the convexity. Catalyst: the same three prints. Invalidation: ISM Services prices paid rising, or payrolls above ~200k on the back of an ISM employment index that just turned positive for the first time in 33 months.

4. Own the airline and cruise complex against energy as the cleanest liquid expression of a sustained lower oil price — but keep it small, because the strait is still closed. Expression: long UAL/AAL/NCLH/CCL against short XLE or the integrated majors, dollar-neutral. Why: Monday’s scoreboard was NCLH +6.64%, UAL +5.82%, CCL +3.33% against Energy −1.26% and the E&Ps down 1–3%. Fuel is 20–30% of an airline’s cost base and the equity beta to a crude move is roughly 1.2× in the opposite direction; the trade paid 7–8 points in a session. Catalyst: any verified resumption of Hormuz transit volume above CBA’s 50% threshold. Invalidation: a re-escalation headline — a strike, a tanker attack, or Iran formally rejecting talks, which its foreign ministry already did once on Monday. Size this as a satellite, not a core position; it is a levered bet on a diplomatic process that one party publicly denies is happening.

5. Fade Korean equity beta against Taiwan; the currency has now confirmed the thesis on both legs of a round trip. Expression: short Kospi/EWY, long Taiex/EWT or TSM, FX-hedged. Why: Friday: Kospi +17.9%, won −1.35%. Monday: Kospi −5.12%, won +0.95%. A market whose currency moves against its own index in both directions is levered domestically, not bought internationally — and Bloomberg supplied the funding evidence Monday. Taiwan is the control: TWD moves with the Taiex. Catalyst: the Korean regulator’s next move on single-stock leveraged ETFs; the Samsung/SK Hynix reaction to AMD and WDC/SNDK this week. Invalidation: USD/KRW below 1,420 on a rising Kospi.

6. Hedge the long book in rates vol and credit, not in equity vol. Expression: CDX HY payers or a 5s30s steepener as the portfolio hedge, funded by selling short-dated S&P upside — not S&P puts. Why: VIX at 15.86 is not expensive, but it is also not the right instrument, because the risk being hedged is a rates-and-credibility risk that has already started moving in the rates-vol complex (MOVE 83.02, +17% in eleven sessions, MOVE/VIX at 5.23×) and has not moved in equity vol at all. IG at 80 bp is one basis point wider than 31 December while CCC is 121 bp wider — the index is not pricing the deterioration the bottom of the stack is showing. Catalyst: CPI on 12 August; the SOFR–IORB spread through mid-August. Invalidation: a July CPI that confirms the energy channel is the whole story, in which case the correct trade is simply to be long risk and this hedge costs carry.

The crowded consensuses to stress-test, with the numbers

1. “The de-escalation is real.” The entire session was built on it. Trump said the U.S. and Iran were in talks to open the strait “literally by tomorrow”; Iran’s foreign ministry said there were no immediate plans for peace talks, and Iran said its Oman channel — not a U.S. channel — was the one making progress. WSJ’s own What’s News line was “Iran Says No Talks With the U.S. on Hormuz Strait.” The market has priced a diplomatic outcome that one of the two parties has publicly denied is occurring. A single tanker attack, a failed transit, or an Iranian statement rejecting the framework re-prices crude by several dollars and takes Monday’s equity gain with it. This is the largest single-headline risk on the tape and it is two-sided within hours, not days.

2. “The de-grossing is over, so the AI complex can only go up.” JPMorgan’s Panigirtzoglou says there is “more limited room for any further deleveraging,” and Monday’s tape agreed emphatically. But the same note observes that elevated short interest in individual technology stocks and in SMH and DRAM “suggests low net exposure to the sector” — a statement about how little dry powder there is on both sides. AMD Tuesday, WDC and SNDK Wednesday, Microchip Thursday. Four semiconductor prints in three days into a complex that has round-tripped 8% twice in a week. The SOX is +1.06% on the day and still 22% below its 14,655.3 52-week high.

3. “Earnings have been so strong that the second half takes care of itself.” BofA: a 77% EPS beat rate, the highest since 2021, with Q2 tracking +45% y/y including Alphabet and Amazon mark-ups and +27% excluding them. Bloomberg: 86% of 307 reporters have beaten on EPS. That is the problem, not the reassurance — the comparison base for the second half now embeds all of it, and roughly 15% of the S&P 500 by market capitalisation reports this week into that base. Raymond James’ Matt Orton framed the test: energy, healthcare, utilities and industrials “benefited from the recent rotation” and “results will help determine whether that relative strength is fundamentally sustainable.” WSJ’s Spencer Jakab is running the other side under the headline “The Buffett Indicator Keeps Flashing Red. Is It Broken?”

4. “Credit is fine.” IG OAS at 80 bp is one basis point wider than 31 December. CCC OAS at 1,006 bp is 121 bp wider. The index-level calm is a composition artefact — the IG index is dominated by issuers who do not need the market. The stress is in the bottom of the stack, in floating-rate loan borrowers, and in a private-credit book that does not mark daily and that Bloomberg’s own Odd Lots is now writing about in the context of insurance balance sheets. Two fund-level accidents were reported on Monday alone (§9d).

5. “Volatility is telling you it’s safe.” VIX 15.86; MOVE 83.02, up 17% in eleven sessions; MOVE/VIX at 5.23× against a long-run 4×. One of those two markets is wrong, and the rates market has both the better information set and the live catalyst calendar.

The two-sided geopolitical tape. Toward de-escalation: the called-off strike; Trump’s “last chance” framing; the Oman transit channel; European gas paring losses; Turkey–Iraq extending a 750,000 b/d pipeline deal. Toward re-escalation: Iran’s public denial of U.S. talks; the UKMTO tanker explosion report off Oman on Sunday; Trump’s own statement that the U.S. Navy effectively “blockades” the strait — “Nothing gets through to Iran, unless we want it to”; continued drone attacks on tankers loading at the CPC terminal on Russia’s Black Sea coast; and the unresolved nuclear question. Separately, Trump’s comment that Exxon and Chevron are “making too much money based on a shortage — I don’t like it” introduces a domestic policy tail (windfall levy, export restriction, or jawboning) that the energy equity market has not priced.

Structural watch items. (i) The funding stack: SOFR above IORB, a 99th percentile at the top of the target range, ON RRP drained to $2.1bn, reserves −$158bn in a fortnight, and a 3-month bill 25 bp over SOFR — with quarter-end on 30 September, two weeks after the FOMC. (ii) Fed credibility and the Jackson Hole reset — Wolfe’s Roth, Citadel Securities’ Shah and RBC’s Calvasina all flagged it on the same day, and it is a term-premium risk no data print can settle. (iii) The yen: a coordinated U.S.–Japan intervention with no BoJ policy change behind it, which every major sell-side desk publicly doubts will hold. (iv) Korea: a 17.9%-then-−5.12% round trip funded by short-term corporate borrowing, with regulators still tightening single-stock leveraged ETF rules. (v) Apple’s supply warning, which has not been retracted and which converts the AI capex boom into a cost-of-goods problem for every hardware franchise that is not building accelerators — AAPL was the only Magnificent Seven name red on Monday. (vi) Pharma consolidation: a $400bn AstraZeneca/Bristol Myers combination would reset antitrust expectations for the entire sector, and the equity market’s Monday verdict — BMY +0.21% after a 5% pre-market spike, AZN −7%, healthcare −0.34% — is that it neither happens nor should.

Source Links and Data Notes & Conflicts are in the companion text file US_CrossAsset_Daily_2026-08-03_DataNotes.txt, saved alongside this report.