Closing Briefing — Wednesday, August 5, 2026

Sources: CNBC market live blog (“Dow jumps more than 250 points for record close and fifth winning day”) plus the CNBC midday and after-hours movers pages, and the ADP, ISM services, Kashkari and Cook items; Bloomberg Markets Wrap (“Oil Holds Three-Day Drop as Iran, Oman Reach Hormuz Agreement”), the Bloomberg /markets and /markets/rates-bonds boards and the HYG/LQD quote pages used in §9; WSJ (“Strong Earnings Power Dow to Another Record”; “Google Overhauls AI Leadership as Longtime Chief Scientist Joins Wave of Exits”; “Negotiators Close In on Deal With Iran to Open Hormuz”) and WSJ Market Data Bonds & Rates (5:04 p.m. ET); Investing.com Major Indices, NDX, SOX and S&P-500-component boards (all stamped 15:59:59); U.S. Treasury Text View; CME FedWatch numeric four-column table (data as of 5 Aug 2026 04:58:17 CT) and Investing.com Fed Rate Monitor (Aug 05, 2026 05:45 PM EDT); FRED ICE BofA OAS series, WRESBAL and RRPONTSYD, and NY Fed SOFR/EFFR reference rates; TradingEconomics commodities and currencies boards and Kitco’s AM/PM metals reports; Finviz Groups (Performance view, rendered); NY Fed August indicator calendar; Earnings Whispers day pages. Full categorized links are in the companion Data Notes file.

1. The failed breakout, and where exactly it failed.

This is the tradable fact of the session. S&P 500 record intraday 7,793.68 → close 7,723.55 (−0.90% from the high, −0.17% on the day). Nasdaq high 26,739.00 → close 26,363.44 (−1.40%). SOX high 12,361.2 → close 12,008.9 (−2.85%). NDX high 29,946.94 → close 29,487.79 (−1.53%). Dow high 54,744.33 → close 54,349.12 (−0.72%). The fade is monotonically increasing in beta: the more semiconductor exposure an index carries, the more it gave back. That is a positioning unwind, not a macro repricing — a macro repricing would have shown up in the curve, and the curve did almost nothing outside the front end (§6). Forward catalyst: Thursday’s initial claims and preliminary productivity and costs at 8:30, then Friday’s payrolls (§7).

2. Two Fed officials called for hikes and the front end rallied anyway.

Kashkari: “Corporate earnings are through the roof… The consumer is hanging in there. The labor market is hanging in there. I look at this constellation and I say, what evidence do I have that monetary policy is particularly restrictive right now? So I argued now is the time to start slowly moving up.” Cook: “Inflation is too high, and I consider the risks to the inflation side of the dual mandate higher than the risks to the employment side at this point. As such, I am prepared to act by raising rates, if necessary.” Both are voters. And the market moved the other way: CME’s September hold rose from 41.6% to 45.6% and the hike fell from 58.4% to 54.4% (§8); the 2-year fell 2 bp. The mechanism is that ADP at 44k against 75k expected, on top of a June print revised down to 95k, is a harder number than a speech — and the ISM services employment index at 47.4 corroborates it. Tradable read: the front end is now trading the labour data and discounting the rhetoric. The invalidation is Friday’s payroll.

3. Alphabet’s 4% break is a governance event, not an earnings event.

GOOGL −4.04% to $362.38 on 42.1m shares; GOOG −4.06% to $360.11 on 28.2m — the worst megacap performance of the session, on no guidance change. The catalyst is organisational: Google reshuffled its AI divisions, chief scientist Jeff Dean is leaving after 27 years to found a startup that Google will invest in, and Demis Hassabis becomes chairman of Google DeepMind (CNBC, WSJ, Bloomberg’s “Google AI Veterans Depart During Seismic Leadership Shift”). The stock traded as high as $384.48 — it was up 1.6% before the story broke and closed 5.75% below that high. For a long/short book the relevant question is whether key-man departures at the research layer lead model quality; the market has just voted that they are worth roughly $95bn of market capitalisation in one session.

4. Gold and silver up 4%+ into hawkish Fed speak.

Spot gold +4.15% to $4,247.27, silver +4.35% to $62.097 (TradingEconomics; Kitco $4,244.00 / $61.880 at its PM cut). Gold miners were the best-performing cohort in the market: GDX +6.8%, Gold Fields +10%, Newmont +6.71% to $104.29, Barrick +6% (CNBC). The conventional read — softer ADP, weaker dollar, lower yields — explains part of it, but the dollar fell only 0.18% and the 10-year did not move. A 4% move on those inputs is not arithmetic; it is a bid for a hedge against a central bank that has three dissenters, a governor volunteering that she would raise rates, a services prices index whose twelve-month average is at a three-year high, and a president who has spent the year attacking it. Second-order tell: gold rallying on hawkishness is the market pricing policy error in either direction.

5. The memory complex is now a two-sided trade — and the after-hours prints went the wrong way.

In the regular session Micron closed exactly unchanged at $892.70 after trading $928.98 (−3.91% from the high), Western Digital −5.36% to $519.17 after trading $564.66 (−8.05% from the high), Seagate −0.89% and Lam Research −3.25%. After the close, SanDisk fell about 4% on Q1 revenue guidance of $10.3–10.8bn against an $11.16bn consensus, despite an $8.79bn quarter versus $8.64bn expected, adjusted EPS of $39.25 and a $14bn buyback increase; Western Digital fell more than 10% on a guide of $4.00 on $4.1bn against $3.81 on $4.04bn (CNBC after-hours). The read: the memory names are beating on the quarter and guiding light on the next one, which is exactly the pattern that converts a momentum trade into a valuation trade. Those reactions belong to Thursday’s tape.

6. Hormuz moved from rumour to draft — and crude did nothing.

Bloomberg’s wrap ran under “Oil Holds Three-Day Drop as Iran, Oman Reach Hormuz Agreement”; WSJ reported a draft that “would give Tehran oversight of ships entering the Persian Gulf but wouldn’t let it levy tolls or service fees”; Reuters sources said the concession on “some form of control over Hormuz” has already been made but pushed back on imminence. Meanwhile Yemen’s Houthis claimed a strike on a Saudi tanker in the Red Sea, which lifted Brent to $80.22 and WTI to $76.12 in the morning. Net: WTI September settled −0.73% at $75.22 and Brent October +0.11% at $79.45. After a 10.4% two-day collapse, a session in which the agreement text emerges and crude moves less than 1% says the deal is now in the price and the residual risk is two-sided.

7. Health care was the best sector on the worst kind of day.

Healthcare +1.19%, the top group, on Eli Lilly +4.76% to $1,168.78 (Q2 beat, 2026 revenue guidance raised on Zepbound and Mounjaro), Amgen +4.57%, McKesson +5.64%, Cencora +3.53%, Centene +4.56%, Elevance +3.50%, UnitedHealth +1.27%, Charles River +11.36% (adjusted EPS and revenue beat FactSet; guidance raised to $11.15–11.45 from $10.80–11.30). Against that, Insulet −20.15%, DaVita −17.24%, DexCom −4.92% and Baxter −3.62%. A group that was −2.96% on the week in Tuesday’s report is now the day’s leader and only −1.30% on the week — the classic late-cycle rotation into defensives dressed up as an earnings story.

8. The AI-power complex could not hold a bid.

Vistra −1.85%, Constellation Energy −1.16%, NRG +3.14% — NRG bounced 3.14% after Tuesday’s −15.48%, recovering less than a fifth of the loss, while the two names that fell less on Tuesday fell again on Wednesday. Utilities −0.96% was the second-worst sector. With Constellation reporting Thursday at 7:05 a.m. and Vistra Friday before the open (§5), the Texas interconnection audit gets its first earnings adjudication in the next 48 hours.

Five green, six red, and the two ends of the table are both single-name stories. Basic Materials +2.71% is gold and copper, not industrial demand: Newmont +6.71%, Freeport-McMoRan +3.11%, Mosaic +3.08%, DuPont +3.01%, IFF +8.88% against Celanese −4.64% and LyondellBasell −3.40% — the metals leg carried the group while the chemicals leg fell, which is a precious-metals bid wearing a cyclical costume. Communication Services −2.35% is Alphabet and essentially nothing else: Finviz buckets Alphabet A (−4.04%) and Alphabet C (−4.06%) here alongside Meta (+0.17%) and Netflix (+0.86%), both green, and Paramount Skydance +4.53% and Walt Disney +3.66% were among the day’s better performers — so a group that fell 2.35% contained four of its largest constituents in positive territory. This is the single most important composition note in the table. Technology −0.29% is a far shallower decline than the SOX’s −1.40% because Finviz’s Technology bucket is not a semiconductor index: Apple +0.51%, Salesforce +1.00%, Intel +0.22% and Monolithic Power +0.61% offset AMD −7.04%, CDW −9.03%, ON Semiconductor −4.79%, Microchip −3.66%, Lam Research −3.25%, Qualcomm −3.16%, Applied Materials −2.26% and Texas Instruments −2.09% — with Nvidia +3.44% doing most of the offsetting work by weight. Industrials −1.51% is the mirror of Tuesday’s +2.71%, giving back 56% of it in one session with Caterpillar −0.62% after trading $905.72 (a second consecutive high-to-close fade, this time 3.83%); note that the iShares US Aerospace & Defense ETF still made an intraday all-time high, led by Kratos +7% on a beat and raised guidance (CNBC), so the weakness was not defence. Energy −1.99% on a flat crude settlement is the tell of the sector table — the beta to the barrel was 2.7x and negative — because the refiners that gained on Tuesday’s widening crack gave it straight back: Marathon Petroleum −4.75%, Valero −2.05%, Phillips 66 −1.62%, alongside EOG −6.47%, Devon −4.47%, Diamondback −2.97%, APA −3.17%, ConocoPhillips −2.46%, Occidental −2.29%, Chevron −2.10%, Exxon −1.51%. Healthcare +1.19% and Financial +0.30% were the only groups doing what a defensive rotation looks like. Source: Finviz Groups (Performance view, rendered); Finviz buckets are not official GICS/S&P sector indices.

Reconciliation. All eleven groups reconcile against Tuesday’s published YTD compounded by Wednesday’s 1-day move: Communication Services 1.0352 × 0.9765 = +1.09% vs. 1.09% shown; Financial 1.0867 × 1.0030 = +9.00%; Consumer Defensive 1.0827 × 1.0017 = +8.45%; Healthcare 1.0462 × 1.0119 = +5.86%; Utilities 1.0328 × 0.9904 = +2.29%; Real Estate 1.1133 × 1.0005 = +11.39% vs. 11.38%; Industrials 1.1627 × 0.9849 = +14.51% vs. 14.53%; Technology 1.2347 × 0.9971 = +23.11% vs. 23.07%; Basic Materials 1.1135 × 1.0271 = +14.37% vs. 14.28%; Consumer Cyclical 0.9888 × 0.9961 = −1.51% vs. −1.42%; Energy 1.3038 × 0.9801 = +27.79% vs. 28.02%. Maximum deviation: 0.23 pt (Energy); median deviation 0.01 pt; five groups exact. Energy’s gap is the largest single-group deviation in a fortnight and follows two sessions of violent producer-versus-refiner dispersion — Finviz’s group aggregation rebalances on market capitalisation, and that is exactly when it shows up. Flagged rather than smoothed.

Up

• Shopify +16.98% to $144.24 on 40.5m shares (not an S&P 500 constituent). The largest one-day move on the composite board, and it was an analyst event ahead of the print, not a print. Morgan Stanley initiated at Overweight with a $192 target (+33.1% from the close); Jefferies upgraded to Buy, $160 (+10.9%); Stifel upgraded to Buy, $150 (+4.0%); Bank of America reinstated Buy, $150 (+4.0%). Traded as high as $153.88 and closed 6.3% below it.

• Charles River Laboratories +11.36% to $260.72. Q2 adjusted EPS and revenue beat the FactSet consensus; full-year adjusted EPS guidance raised to $11.15–11.45 from $10.80–11.30.

• IFF +8.88% to $88.07, Assurant +7.20% to $301.57 — both post-print, both Tuesday-after-close reporters.

• Newmont +6.71% to $104.29 on 8.3m shares, part of the broadest precious-metals move of the year: GDX +6.8%, Gold Fields ADR ~+10%, Barrick ~+6% (CNBC midday).

• Booking Holdings +6.56% to $207.02, Marriott +4.67%, Hilton +4.04%, Wynn Resorts +3.67%, Expedia +2.44% — the travel complex was the cleanest expression of cheaper jet fuel and a softer dollar, and Wynn’s $109.24 high against a $101.18 close (−7.38%) was the largest fade in the group.

• McKesson +5.64% to $877.23 and Cencora +3.53% — the drug distributors, reporting after the close.

• Eli Lilly +4.76% to $1,168.78 on 5.5m shares. Q2 EPS and revenue beat; 2026 revenue guidance raised on Zepbound and Mounjaro demand. Traded to $1,216.94; closed 3.96% below the high, the largest megacap fade of the day.

• Amgen +4.57% to $407.85, Centene +4.56%, Elevance +3.50%, IQVIA +3.44%, Revvity +3.26%, Danaher +2.49%, Thermo Fisher +2.32% — the health-care breadth behind §2 item 7.

• Walt Disney +3.66% to $101.77 on 18.0m shares on a mixed fiscal Q3 (EPS beat, revenue slightly light; experiences revenue +10% y/y). Morgan Stanley’s Sean Diffley — “There’s lots to like” — keeps Overweight and a $123 target, 25% above Tuesday’s close and 20.9% above Wednesday’s; he flags that reiterated full-year guidance “underscores the durability and diversification across the business” and that the sale of the A+E stake to Hearst shows new CEO Josh D’Amaro will “simplify & optimize with a clear move away from linear.” Wells Fargo (Overweight, Steven Cahall) and Bank of America (Buy) both reiterated.

• NVIDIA +3.44% to $219.23 on 138.2m shares — the single largest positive contributor to the Dow. Catalyst: Elon Musk on SpaceX’s earnings call — Nvidia has “the best AI computer,” and “we’re exclusive to Nvidia.”

• Arista Networks +3.57% to $197.31 — but it traded $214.89 intraday and closed 8.18% below that high, the widest high-to-close fade of any large-cap gainer on the board.

• Fair Isaac +4.26%, Yum! Brands +3.18%, BorgWarner +5.93%, Paramount Skydance +4.53%, NRG +3.14% to $120.72 (recovering less than a fifth of Tuesday’s −15.48%).

• Shake Shack +12.25% to $74.33 (not an S&P 500 constituent) after Starboard Value’s Jeff Smith told Bloomberg the firm has built a position of “hundreds of millions of dollars” and is likely the largest active shareholder.

Down

• Insulet −20.15% to $133.21 on 6.8m shares — the worst S&P 500 performer. The company lowered full-year 2026 revenue growth guidance and cut its U.S. Omnipod sales outlook.

• DaVita −17.24% to $188.69. Beat estimates but reported declining revenue per treatment and rising patient-care costs, and only reaffirmed rather than raised full-year guidance. Traded $208.53 to $183.62.

• SpaceX −13.61% to $108.27 (not an S&P 500 constituent). First quarterly report since the June listing: Q2 capital expenditure up sixfold to $18.4bn, ahead of expectations, with the majority going to AI. Bloomberg: “SpaceX’s $101 Billion Unlock Heaps Pressure on Battered Shares.”

• CDW −9.03% to $140.10. Margin squeeze plus an announced executive departure; the intraday low of $113.00 against a $143.67 high is a 21.3% range, the widest of any S&P name.

• AMD −7.04% to $482.05 on 47.4m shares. Adjusted earnings only slightly ahead of consensus on record sales — the second consecutive session in which a semiconductor beat was sold.

• Match Group −7.49%, Enphase −6.75%, EOG Resources −6.47% to $134.23 (post-print, traded $144.02).

• Western Digital −5.36% to $519.17 into an after-hours guide that took it down a further ~10%; SanDisk fell ~4% after hours.

• Uber −5.31% to $68.17 on 46.2m shares. Q2 EPS of $0.81 in line; revenue $14.19bn vs. $14.24bn (LSEG); Q3 bookings guided to $59.25bn at the midpoint vs. $59.33bn (StreetAccount) and Q3 EPS to $0.84–0.88 vs. $0.89. A clean guide-down.

• CVS Health −5.08% to $99.12 on 18.2m shares — and this is the fade of the day. CVS beat on both lines (EPS $2.58 vs. $1.85 FactSet; revenue $106.1bn vs. $100.03bn) and raised adjusted EPS guidance to $7.90–8.10 from $7.30–7.50, and was up around 3% pre-market. It closed down 5.08%, having traded as low as $93.50. A 20%+ EPS beat with raised guidance that closes 5% lower is the most informative single print of the session about how much good news is already in managed-care and pharmacy multiples.

• SBA Communications −5.06%, Crown Castle −4.92%, American Tower −4.10% — the tower REITs were the worst sub-industry in the market on a day the 10-year did not move, which removes the usual rates explanation and leaves the data-centre/AI-adjacent capital-allocation debate.

• ON Semiconductor −4.79%, Microchip −3.66%, Teradyne −3.51%, Lam Research −3.25%, Qualcomm −3.16%, First Solar −2.80%, Applied Materials −2.26%, Texas Instruments −2.09%.

• Marathon Petroleum −4.75%, Devon −4.47%, Diamondback −2.97%, APA −3.17%, ConocoPhillips −2.46%, Occidental −2.29%, Chevron −2.10%, Valero −2.05%.

• Alphabet C −4.06% / Alphabet A −4.04% (§2 item 3). Super Micro −4.32%, Celanese −4.64%, DexCom −4.92%, Baxter −3.62%, Bristol-Myers −3.43%, Kraft Heinz −3.42%, LyondellBasell −3.40%, Edison International −3.35%, NiSource −3.64%, Fidelity National Information Services −3.21%.

• Flutter Entertainment −5%+ (not an S&P 500 constituent) after CEO Peter Jackson’s departure was announced, with international head Dan Taylor succeeding him on 1 October, alongside lowered full-year revenue guidance.

Times are ET; S&P 500 components only. Every day page from Thursday 8/6 through Friday 8/14 was re-pulled from Earnings Whispers this session (`/1` = before open, `/2` = after close) and screened against the current S&P 500 constituent list. Re-verify times and membership against company IR before trading any date.

Mon 8/3 — completed. BMO: Loews (L), Marriott (MAR), Tyson (TSN). AMC: SBA Communications (SBAC), Vertex (VRTX), Diamondback (FANG), Palantir (PLTR), ON Semiconductor (ON), Alexandria Real Estate (ARE), Clorox (CLX), ONEOK (OKE), Williams (WMB).

Tue 8/4 — completed. BMO: ADM, Ball (BALL), DuPont (DD), Gartner (IT), Henry Schein (HSIC), Leidos (LDOS), Revvity (RVTY), Waters (WAT), Apollo (APO), Caterpillar (CAT), IDEXX (IDXX), Kimberly-Clark (KMB), Merck (MRK), Zebra (ZBRA), Aptiv (APTV), Marathon Petroleum (MPC), Pfizer (PFE), Kimco (KIM), AMETEK (AME), Broadridge (BR), Duke (DUK), McDonald’s (MCD), NRG, Rockwell (ROK), TransDigm (TDG), Cummins (CMI), FIS, PSEG (PEG), Sysco (SYY), W.W. Grainger (GWW), Progressive (PGR), Expeditors (EXPD), Pinnacle West (PNW). AMC: Amgen (AMGN) — closed +4.57% Wednesday, Booking (BKNG) — +6.56%, Gilead (GILD), Wynn (WYNN) — +3.67%, Arista (ANET) — +3.57%, DaVita (DVA) — −17.24%, Devon (DVN) — −4.47%, Emerson (EMR) — +2.29%, Jacobs (J), Match (MTCH) — −7.49%, AMD — −7.04%, Celanese (CE) — −4.64%, Healthpeak (DOC), IFF — +8.88%, Mosaic (MOS) — +3.08%, Prudential (PRU), Assurant (AIZ) — +7.20%.

Wed 8/5 — completed. BMO: EOG Resources (EOG) — −6.47%, Cencora (COR) — +3.53%, CVS Health (CVS) — beat and raised, closed −5.08%, NiSource (NI) — −3.64%, Zimmer Biomet (ZBH) — +2.45%, Eli Lilly (LLY) — beat and raised, +4.76%, Iron Mountain (IRM), Global Payments (GPN), Uber (UBER) — −5.31% on soft Q3 guidance, CDW — −9.03%, Charles River Labs (CRL) — +11.36%, Insulet (PODD) — −20.15%, the day’s worst S&P performer, Kraft Heinz (KHC) — −3.42%, Phillips 66 (PSX) — −1.62%, Honeywell Aerospace (HONA), Walt Disney (DIS) — +3.66%; Morgan Stanley OW, $123 PT. AMC: Western Digital (WDC) — −5.36% in the session, a further ~10% after hours on soft guidance, SanDisk (SNDK) — ~−4% after hours despite an $8.79bn quarter and a $14bn buyback increase, Axon (AXON), Expedia (EXPE), AppLovin (APP), Block (XYZ), Corpay (CPAY), DoorDash (DASH), eBay (EBAY), Realty Income (O), Solventum (SOLV), McKesson (MCK) — +5.64% into the print, Motorola Solutions (MSI), Albemarle (ALB), MetLife (MET) — Bloomberg: “MetLife’s Earnings Beat Wall Street Expectations”, News Corp B (NWS), Occidental (OXY), Texas Pacific Land (TPL), News Corp A (NWSA), CF Industries (CF), Host Hotels (HST), Steris (STE), Atmos Energy (ATO), Allstate (ALL).

Thu 8/6 — re-pulled this session (both buckets); initial claims and preliminary productivity & costs at 8:30. BMO: Targa Resources (TRGP) 6:00, Becton Dickinson (BDX) 6:30, Kenvue (KVUE) 6:30, Molson Coors (TAP) 6:30, Viatris (VTRS) 6:55, ConocoPhillips (COP) 7:00, Datadog (DDOG) 7:00, Evergy (EVRG) 7:00, Fiserv (FISV) 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 8:00, Fox Class B (FOX) 8:00, Ralph Lauren (RL) 8:00, Fox Corporation (FOXA) — listed “Before Open”; the reviewed calendar did not publish a specific time, so confirm with Fox IR. CEG at 7:05 is the most important print of the week for the AI-power complex given the Texas interconnection audit, and COP at 7:00 is the first major integrated to report into a post-Hormuz crude curve. AMC: Airbnb (ABNB) 4:00, Akamai (AKAM) 4:00, The Trade Desk (TTD) 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 4:15, Microchip Technology (MCHP) 4:15, Consolidated Edison (ED) 4:30.

Fri 8/7 — re-pulled this session (both buckets); payrolls day. BMO: Vistra (VST) — listed “Before Open”; no specific time published, so confirm with Vistra IR — the print that attaches an earnings number to the Texas moratorium. Take-Two Interactive (TTWO) 7:00, PPL 7:30. AMC: the reviewed after-close page for 8/7 lists three names (HE, GLBS, GLXZ), none an S&P 500 constituent — no S&P 500 after-close reporters are published for that date, unchanged for a fifth consecutive edition.

Mon 8/10. BMO: the reviewed before-open page lists no S&P 500 reporters (50 names screened). AMC: Simon Property Group (SPG) 4:05.

Tue 8/11. BMO: Cardinal Health (CAH) 6:45. AMC: Super Micro Computer (SMCI) 4:05, Lumentum (LITE) 4:00.

Wed 8/12 — CPI day (8:30). BMO: Amcor (AMCR) 6:00, Trimble (TRMB) 6:55 (new to this pull). AMC: Cisco (CSCO) 4:05, Coherent (COHR) 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, business inventories and Michigan preliminary. Neither page lists an S&P 500 reporter (nine names screened across both).

Changes vs. the prior calendar (8/4 report). Addition: Trimble (TRMB) 6:55 on 8/12 BMO — the only genuine new S&P 500 entry across the ten day pages re-pulled. No removals: every 8/6 and 8/7 name carried by the prior edition re-appeared with an identical timestamp, including the three previously flagged as absent-then-returned (DDOG, FISV, TTD) — a second consecutive confirmation. Ferguson Enterprises (FERG) 6:45 appeared on the 8/10 before-open page and is excluded here for consistency with the prior edition’s screen of that same page; it is a borderline membership call, listed in the companion Data Notes, so confirm with company IR if it matters to a position. Timing buckets still unpublished: FOXA (8/6 BMO) and VST (8/7 BMO) are listed “Before Open” with no clock time — confirm with Fox IR and Vistra IR respectively. The forward calendar remains almost empty: eleven S&P 500 reporters over 8/10–8/14 against roughly 130 this week; Cisco (8/12 AMC) and Applied Materials (8/13 AMC) are the only two that move an index, and both land on inflation-print days. Conservatively excluded non-constituents are listed in the companion Data Notes.

The bill curve is the separate story, and it is a financing signal. The 3-month is +6 bp on the week while the 10-year is −4 bp — a 10 bp flattening of 3M10Y in five sessions, to 74 bp from 84 bp. Monday’s 8 bp cheapening in the 3-month bill did not fully retrace; it has settled in as a new level, and the 1-month is +4 bp on the week. This is not a policy signal — the 6-month and 1-year are both flat-to-lower on the week, which they could not be if the market were pricing more near-term tightening. It is a supply-and-funding signal, and §9 pairs it with reserve balances at $2.985tn, an ON RRP at $1.65bn and a SOFR–IORB spread that has flipped positive.

Real-time versus official par. WSJ’s 5:04 p.m. marks read 10-year 4.617% (−0.2 bp), 30-year 5.171% (−0.4 bp), 7-year 4.467% (0.0), 5-year 4.334% (+0.1), 3-month bill 3.680% (−0.5 bp); Bloomberg’s board shows the 10-year at 4.61. The official par 10-year of 4.63% against a 4.617% real-time close is a 1.3 bp baseline-and-timing artefact — Treasury’s par yields are struck at 3:30 p.m. off a bid-side curve, WSJ’s at 5:04 p.m. off dealer quotes — not a disagreement about the level, and the directions agree everywhere. Note also that WSJ’s 3-month bill at 3.680% sits 21 bp below the official par 3-month of 3.89%: those are different instruments (a discount-basis secondary-market bill quote versus a coupon-equivalent par yield), and the gap is the conversion plus the on-the-run basis, not an error in either.

Current week — released

Current week — remaining

Next week

Current target range: 3.50–3.75% (held 9–3 on 29 July; Hammack, Kashkari and Logan dissented for +25 bp; IORB 3.65%). Chair: Kevin Warsh. Next FOMC: Wednesday, September 16, 2026, 2:00 PM ET — 42 days away. CME FedWatch September card read live: contract ZQU6, expiry 30 Sep 2026, mid price 96.3025, prior volume 52,956, prior open interest 232,684, data as of 5 Aug 2026 04:58:17 CT. Investing.com Fed Rate Monitor cards stamped Aug 05, 2026 05:45 PM EDT.

CME FedWatch headline — September 16, 2026 meeting

Column sums: 100.0 / 100.0 / 100.0 / 100.0 — every column sums exactly; no rounding residual to disclose this session.

Provenance of every column. CME published its complete numeric four-column table again today, dated 1 DAY = 4 Aug 2026, 1 WEEK = 29 Jul 2026, 1 MONTH = 2 Jul 2026, so all four columns are read directly off CME; none is carried, estimated or reconstructed. The validation matters: the 1-DAY column reads 41.6% hold / 58.4% hike, which is precisely what this report published as NOW on Tuesday. That is the second consecutive session in which a live read reconciles to CME’s subsequent end-of-day settlement snapshot to the tenth of a point. Note that the 1-WEEK reference date has rolled from 28 Jul to 29 Jul, so the 58.3% cumulative shown here is not comparable with the 76.0% the prior edition printed for 28 Jul; 29 July was FOMC day and the distribution collapsed after the statement. Standing caveat: CME’s historical columns are end-of-day settlement snapshots, and a live read taken after ~5:00 p.m. ET is indicative only.

Reconciling CME against Investing.com — the gap is 2.0 points and it decomposes cleanly. For September, CME puts the hold at 45.6% and the +25 bp at 54.4%; Investing.com puts them at 47.6% and 52.4%. CME’s mid price is 96.3025 and Investing.com’s future price is 96.305 — a difference of 0.25 bp of contract price, with Investing.com quoting the higher (more dovish) price. On the ~3-points-of-headline-probability-per-basis-point sensitivity this report calibrated in late July, 0.25 bp of price accounts for roughly 0.75 points of the gap; the residual ~1.25 points is methodology — CME day-weights the September contract (the meeting is 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, with EFFR fixing at 3.63%, while Investing.com maps price to outcomes on a simpler basis. A ~1.25-point methodology residual is within the ~1-point-plus band observed all week. Both vendors agree on direction and on the sign of every change. 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 100.1 / 100.0 / 100.1; December 100.1 / 100.0 / 100.0 — the current October and December columns are the two cells that round to 100.1 this session. No easing is priced at any 2026 meeting in any column — 0.0% below 3.50% throughout, for the thirteenth consecutive session. Contract prices behind the cards: Sep 96.305, Oct 96.225, Dec 96.090 (implied 3.695%, 3.775%, 3.910%).

Multi-day momentum — a fourth consecutive session of dovish drift, and the front card is now doing most of the work. On CME’s cumulative basis the September hike arc reads 54.1% a month ago → 58.3% a week ago (29 Jul, post-FOMC) → 58.4% a day ago → 54.4% now: down 4.0 points on the day, down 3.9 points on the week, up 0.3 points on the month. On Investing.com’s basis the September hold has gone 37.9% → 42.3% → 47.6%: +9.7 points in five sessions and +5.3 points today. The back of the 2026 curve is drifting the same way but far more slowly: October cumulative 73.5% → 68.4% → 65.6%; December cumulative 83.8% → 81.6% → 80.1%. The shape of the drift is the point. September has moved 9.7 points in a week; December has moved 3.7. The market is not removing hikes — it is postponing the first one. December’s +25 bp bucket is 44.1% against 43.9% a week ago — statistically unchanged — while its hold has risen from 16.2% to 20.0% and its ≥+50 bp cohort has fallen from 39.9% to 36.0%. Read together: the modal outcome is still one hike by year-end; what has changed is that the market no longer thinks it arrives in September.

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

The terminal rate the strip draws is 4.04%, reached in June–July 2027, and it has fallen ~3 bp in a session and ~7 bp in a week. The contract prices are the cleanest way to see it: 95.960 across both the June and July 2027 meetings — a flat trough rather than a point peak — rising to 96.040 by December 2027, so the market prices a peak of 4.04% and about 8 bp of give-back by end-2027, less than half a cut off the peak across seventeen months. Two structural changes today. First, the modal range at the April 2027 meeting has flipped back down to 3.75–4.00 from 4.00–4.25, so 4.00–4.25 is now modal at only one meeting on the entire strip (June 2027, 34.0%) — a week ago it was modal at four. Second, the easing tail is thickening at the back: 6.5% below 3.50% by December 2027 against 5.2% on Tuesday, and 3.4% by October 2027 against 2.8%. Cumulative tightening probability peaks at 87.8% in June 2027 — the market remains ~88% certain the Fed is higher than today at some point in the next year, and ~93.5% certain it is not lower by end-2027.

(3) Year-end probability ladders

Ladder sums: 2026 = 100.1 (current), 100.0 (prev-day), 100.0 (prev-week); 2027 = 99.9 (current) — the December-2027 card rounds one-tenth light, as do the July and October 2027 cards, while the January, March, April and June 2027 cards sum to exactly 100.0 and the September 2027 card to 100.1. Stating it explicitly because the ladder invites the question: the probability of any 2026 rate cut is 0.0%, and the probability of a cumulative 100 bp of tightening by end-2026 is also 0.0%.

(4) Rate-path interpretation

One day. The September hold rose 4.0 points on CME (41.6% → 45.6%) and 5.3 points on Investing.com (42.3% → 47.6%), the largest single-session dovish move on the September card since 29 July. Every 2026 and 2027 card added to its hold bucket: October +2.9, December +1.6, January-2027 +1.8, March-2027 +1.7, April-2027 +1.6, June-2027 +1.5, July-2027 +2.0, September-2027 +1.4, October-2027 +2.2, December-2027 +2.3 points. The distribution of the move is the diagnostic: it is roughly five times larger at September than at December, and December’s +25 bp bucket actually fell 0.5 points while its hold rose 1.6. The market did not remove a hike today. It moved one. The mechanism is straightforward: ADP at 44k against 75k, on a June revised down to 95k, with ISM services employment at 47.4 — three labour datapoints that speak to the September window specifically, delivered on a day when Kashkari argued for hikes on CNBC and Cook said she is “prepared to act by raising rates.” The front end sided with the data.

One week. From 29 July (FOMC day), the September hike is down 3.9 points on CME (58.3% → 54.4%) and 9.7 points on Investing.com’s hold-equivalent basis; the October ≥+50 bp bucket is down 4.1 points (18.6% → 14.5%); December ≥+50 bp is down 3.9 points (39.9% → 36.0%). And this is where the cross-asset tension lives: the 30-year Treasury is 3 bp lower on the week and the 20-year 3 bp lower, so unlike last week the long end is no longer diverging from the front end — for the first time in a fortnight the whole coupon curve moved the same direction. But the 3-month bill is +6 bp on the week, which means the divergence has migrated from the long end to the bill end (§6, §9). A week ago the story was “the front end prices out hikes while the long end demands term premium.” This week the story is “the whole coupon curve rallies while the bill curve cheapens.” The first is a policy-versus-term-premium argument; the second is a funding argument, and funding arguments resolve faster.

The named macro hooks, in order of when they can move the card. (i) Thursday 8/6, 8:30 — initial claims, now the tie-breaker after ADP. (ii) Thursday 8/6, 8:30 — Q2 preliminary unit labour costs, the only release this week that can reconcile a 47.4 employment index with a 70.3 prices index. (iii) Friday 8/7, 8:30 — July payrolls, against a June print of just 57k; a sub-50k print would be the fourth consecutive soft labour reading and takes the September hold above 55%. (iv) Friday 8/7, 11:00 — NY Fed Survey of Consumer Expectations, materially more interesting than usual with ISM prices at 70.3. (v) Wednesday 8/12, 8:30 — July CPI, still the only release before 16 September that can move the card by more than 10 points alone. Against that stand the two speeches: Kashkari (“now is the time to start slowly moving up as we get more data in”) and Cook (“I am prepared to act by raising rates, if necessary”), which together with the three July dissents mean the committee’s hawkish wing is now at least four voices. The futures distribution is not mispricing that split — a 45.6/54.4 September card is a precise description of a committee that could plausibly go either way.

(a) IG and HY credit spreads

FRED publishes the ICE BofA OAS series with a one-business-day lag. The table below is as of Tuesday 4 August 2026 — it is not a same-day mark. Same-day direction is cross-checked against the cash market and against Bloomberg/WSJ credit coverage beneath the table.

CDX retrieval note — the six-step ladder was worked again and the level remains publicly unobtainable. (1) Bloomberg in Chrome: `/markets/rates-bonds` publishes the Bloomberg Fixed Income Indices (Global Aggregate 500.32, U.S. Aggregate 2,349.35, Asian-Pacific Aggregate 192.03, Pan-Euro Aggregate 227.99, EM USD Aggregate 1,407.39) and the sovereign 10-year grid, but no CDX quote; the Markets Wrap carried no credit-index level. (2) WSJ Market Data Bonds & Rates: Treasurys, money rates, consumer rates and a sovereign grid, no CDX. (3) ICE / S&P Global index pages: product and methodology pages only; levels behind entitlement. (4) FT Markets Data / Reuters credit wraps: no CDX level in the reviewed material for 5 August. (5) cbonds: CDX.NA.IG 5Y and CDX.NA.HY 5Y are carried but masked (“* bps”), with the last public reference now shown as 31/07/2026 — one day older than the reference cbonds displayed in the prior edition, so the public series is going stale rather than updating. (6) Cash-market proxy, labelled as such: HYG closed −0.03% at $79.52 and LQD −0.01% at $106.74. The proxy read is unusually informative today precisely because it is so flat: on a session in which the S&P failed at a record and the Nasdaq fell 0.83%, neither credit ETF moved more than three cents. Credit did not participate in the equity fade at all. All six steps are named so the gap is auditable rather than asserted. Quoting convention: CDX IG 5y is quoted in basis points of spread and CDX HY 5y in price points, where a rising price means tightening credit spreads.

The direction cross-check. The IG credit spread at 78 bp is 1 bp inside where it started the year and 3 bp tighter on the week; the HY credit spread at 273 bp is 8 bp inside its 31 December level and 11 bp tighter on the week — the tightest HY print of the past fortnight. The one series still not participating is the low-quality tail: the CCC credit spread at 1,019 bp is 14 bp wider on the week and 134 bp wider year-to-date. The CCC-minus-HY differential is 746 bp against 604 bp at the start of the year — that gap has done all the work of pricing this cycle while the index has priced none of it, and it narrowed by 4 bp on 4 August, the first meaningful compression in a week.

(b) Money-market and funding plumbing

(c) Rates volatility and swap spreads

MOVE has now fallen for two consecutive sessions, from 83.02 to 77.56, and is nonetheless up 18.6% on the month. VIX fell 4.18% on Wednesday. For the first time in a fortnight, rates volatility and equity volatility are falling together — and they are doing it on a day when ADP missed by 41%, ISM services employment hit a five-month low, two Fed officials publicly argued for hikes and the S&P failed at a record intraday high. That is a remarkable amount of information for two volatility markets to absorb by going down. The MOVE/VIX ratio at 4.91× is essentially unchanged from Tuesday’s 4.88×, so the relative pricing has not moved; both simply got cheaper. The Investing.com MOVE card again publishes an internally inconsistent “Prev. Close” of 95.74 against an intraday range of 77.56–80.48 and a 52-week range of 55.77–115.02; that field is withheld here rather than quoted. The same caution applies to the VIX board’s 18.43 session high, which sits 2.62 points above the close, carries a 16:14:46 stamp — after the 4:15 p.m. VIX settlement — and could not be corroborated against a second vendor, so no interpretation is drawn from it.

(d) Issuance, leveraged loans and private credit

• The primary calendar produced a live data-centre financing today: Celestica joined the data-centre funding rush with a $3bn deal (Bloomberg). That is the theme this report flagged on Tuesday arriving as an actual new-issue print rather than a forecast.

• IG primary remains on pace to challenge all-time monthly issuance records, with gross IG supply projected above $2tn for 2026 against $1.7tn in 2025.

• HY primary ran roughly $5.7bn in the most recent reported week, with sentiment softening on rate volatility and data-centre supply concerns.

• Six of the largest hyperscalers have issued more than $150bn of publicly traded debt in 2026, with more expected. This remains the single largest structural change in the IG index’s composition this cycle, and the Texas interconnection audit is therefore a credit story as much as an equity one — Constellation Energy reports Thursday 7:05 a.m. and Vistra Friday before the open (§5).

• Named private-credit and single-name watch items. “Hedge Funds Take Big Hit in July After Bruising AI Selloff” (Bloomberg) — the July drawdown is now a reported number, not an anecdote, and follows the levered AI fund forced to liquidate into lender redemptions reported last week. Jefferies has been told that some of the invoices underpinning its financing to iron-ore trader Sapphire Minmetals Corp. are not genuine (Bloomberg). Cerberus is seeking $4bn for a second supply-chain fund (Bloomberg).

• Morningstar LSTA leveraged loan index: No reliable data available at this time — the index level was not published in any of the reviewed sources for 5 August. Bank CDS falls under the CDX retrieval gap documented above.

Quote basis: all pairs are spot in the market convention shown; a positive %Chg on a USD/XXX pair means the dollar strengthened, and on EUR/USD, GBP/USD and AUD/USD it means the dollar weakened. EUR/USD, USD/JPY, USD/CNY, USD/CHF and DXY carried a rolling 06:18 intraday stamp rather than an Aug/05 daily stamp at the time of the pull and may differ from a 4 p.m. New York mark on timing alone.

Basis caveats. WTI and Brent above are CNBC-reported settlements for the September and October contracts respectively; the TradingEconomics rows are a rolling front-month quote stamped 06:17 the following morning and are shown separately rather than blended. The 15-cent WTI gap (0.2%) is well inside the ~1% withholding threshold this report applies to TradingEconomics crude. Gold is quoted spot; Bloomberg’s board shows $4,307.60 and WSJ’s $4,306.60, both futures marks — a $60 basis, or 1.4%, which is the carry on the December contract at a 4.3% funding rate, not a vendor error. Copper’s $6.7275 is a September settlement (CNBC); the TradingEconomics $6.7215 is spot, and the 0.6-cent gap is the spot-futures basis.

Desk-style expressions for institutional accounts. Each carries an explicit catalyst, invalidation and sizing note. This is not personalized investment advice — verify independently before acting.

1. Long the September FOMC hold, funded in October (long ZQU6 / short ZQV6).

Expression: buy the September hold via ZQU6, sell ZQV6 to fund. Catalyst: July payrolls Friday 8/7 8:30, against a June print of 57k and an ADP of 44k; initial claims Thursday. Rationale: September is 45.6/54.4 on CME while October cumulative is 65.6%, so the strip prices roughly 11 points of “delayed, not cancelled.” A soft payroll compresses September far faster than October. Invalidation: a payroll above 150k or a falling unemployment rate. Sizing: a two-day event trade into an 8:30 print; size for the gap, not for carry.

2. Long gold miners against the metal, or long gold outright on a pullback.

Expression: long GDX or a basket (Newmont, Barrick, Gold Fields) against a partial short in gold futures, or simply long spot on any retracement toward $4,150. Catalyst: CPI 8/12; the NY Fed Survey of Consumer Expectations Friday 11:00; any further Fed dissent. Rationale: gold is −1.67% year-to-date after a 4.15% day, so this is a rally off a drawdown with an under-positioned momentum cohort, and miners have operating leverage to a spot price 26% higher year-on-year. Invalidation: a payroll above 150k plus a benign CPI, which would restore the “Fed is credible and will simply hike” narrative and take gold back through $4,050. Sizing: miners carry roughly 2x the metal’s beta; size the equity leg at half the notional you would use in the metal.

3. Fade the crack-spread trade; own distillate over gasoline.

Expression: long heating oil versus gasoline, or in equity form long the integrateds (Exxon, Chevron) against the pure refiners (Marathon Petroleum, Valero). Catalyst: the Hormuz signing, ConocoPhillips’ 7:00 print Thursday, weekly EIA inventories. Rationale: heating oil is +14.43% on the month against gasoline −5.79%; the refiners’ Tuesday gain reversed in full on Wednesday, which tells you the crack was a two-day trade rather than a structural widening. Invalidation: a Hormuz deal that actually moves Gulf distillate cargoes within weeks rather than months would collapse the distillate premium. Sizing: dollar-neutral; the beta differential between integrateds and refiners is small enough that beta-neutrality adds little.

4. Short the tower REITs against the equity-REIT complex.

Expression: short SBAC/CCI/AMT against long a diversified REIT basket or the sector ETF. Catalyst: Constellation Energy Thursday 7:05, Vistra Friday before the open, and any further state action on data-centre interconnection. Rationale: SBAC −5.06%, CCI −4.92% and AMT −4.10% on a day the 10-year did not move at all removes the rates explanation entirely; something is repricing digital-infrastructure capital structures, and it is not visible in the 78 bp IG credit spread (§9). Invalidation: a 15 bp-plus rally in the 10-year, which would lift the whole complex and swamp the idiosyncratic signal. Sizing: the towers carry roughly 1.3x the sector’s duration beta; size the long leg up accordingly to stay duration-neutral.

5. Own volatility into 8/12 CPI, not into the payroll.

Expression: buy VIX call spreads or S&P puts dated past 12 August. Catalyst: July CPI, 8/12 8:30. Rationale: VIX fell 4.18% to 15.81 on a session in which the index failed at a record high, ADP missed by 41% and two Fed officials called for hikes. The MOVE/VIX ratio at 4.91× says rates vol is already priced for uncertainty and equity vol is not. Invalidation: VIX through 14, which would signal that the dip-buying reflex is genuinely self-reinforcing rather than complacent. Sizing: own convexity, not delta; a 15.81 VIX makes the option cheaper than the view.

The crowded consensuses, with the numbers attached

1. “The AI capex cycle is a one-way trade.”

AMD fell 7.04% on a beat. SpaceX fell 13.61% on capex that beat expectations. Western Digital fell 5.36% and then ~10% more after hours on guidance. SanDisk fell ~4% after hours despite an $8.79bn quarter, a beat on EPS and a $14bn buyback increase. Alphabet fell 4.04% on an internal reorganisation. Five different flavours of good news being sold in a single session. Against that, Nvidia rose 3.44% on a customer’s verbal exclusivity commitment. The consensus is not wrong about the cycle; it is wrong about the price already paid for it. Stress test: what does the SOX do if Micron’s next guide is in line rather than above?

2. “The Fed’s next move is up.”

CME prices a 54.4% September hike and 0.0% chance of any 2026 cut — for a thirteenth consecutive session. But the September hold has risen 9.7 points in five sessions on Investing.com’s basis, ADP printed 44k, June was revised to 95k, June payrolls were 57k and ISM services employment is at 47.4. Stress test: if Friday prints below 50k, the 0.0% cut probability across every 2026 meeting becomes the most obviously mispriced number in this report.

3. “Credit is fine.”

IG credit spreads at 78 bp are 1 bp inside January; HY at 273 bp is 8 bp inside December; HYG and LQD moved less than three cents on a failed-breakout day. The CCC credit spread is 134 bp wider year-to-date and the CCC-minus-HY differential is 746 bp against 604 bp in January. Stress test: the index is priced for a world in which only IG issuers exist.

4. “Equity vol is correctly priced.”

VIX 15.81, −4.18%, on a record-high failure. MOVE is +18.6% on the month. Stress test: the ratio has been climbing for a month on the numerator, and a MOVE/VIX above 5 has historically resolved by equity vol rising to meet rates vol rather than the reverse.

Source Links and the full Data Notes & Conflicts section are in the companion plain-text file US_CrossAsset_Daily_2026-08-05_DataNotes.txt.