Closing Briefing — Wednesday, August 12, 2026

Sources: CNBC market live blog for 12 August ("S&P 500 closes higher after tame consumer inflation report, tech sector rises"; "July CPI matches economist expectations"; "Stocks making big moves: Wendy's, National Vision, Aecom"; "Nebius, Cava, Super Micro Computer among the names making moves before the bell"; "Cerebras, Coherent among the stocks making big moves after the bell"; "The market is 'gravitating towards December' for a rate hike, Sage's Rob Williams says"; "Goldman reiterates Buy rating for Dell, HPE and NetApp"; "Hormuz closure squeezes global economy as oil demand destruction intensifies, IEA says"); Bloomberg /markets and /markets/rates-bonds boards, "Nebius Reports 514% Jump in AI Cloud Sales as Demand Booms", "Cisco's AI Outlook Disappoints Investors Seeking Bigger Payoff", "CoreWeave Warns of Difficulty If It Must Shift From Nvidia Chips" and "Investors Abandon Hedges in Order to Chase Soaring Stocks"; WSJ Market Data Bonds & Rates (8:56 PM ET); Investing.com Major Indices, NDX, SOX, Russell 2000, U.S.-500-component, Dow-component and European-indices boards (equity boards stamped 15:59:58–15:59:59); U.S. Treasury Text View (official par curve, 12 Aug); CME FedWatch (complete numeric four-column table, "Data as of 12 Aug 2026 07:48:00 CT") and Investing.com Fed Rate Monitor (Aug 12, 2026 08:35–08:45 PM EDT); FRED ICE BofA OAS series, IORB, WRESBAL, RRPONTSYD; NY Fed reference-rates API and the August indicator calendar; TradingEconomics commodities and currencies boards; Finviz Groups (Performance table view, rendered); Earnings Whispers day pages 8/13–8/21. Links in the Source Links appendix.

1. The AI-hardware trade re-rated in one session, and the money came out of megacap software. Super Micro +19.02% to $37.61 on first-quarter guidance of $1.01–1.10 adjusted EPS against a $0.76 LSEG consensus and revenue of $14.5–15.5bn against $11.68bn — a 25% beat on the revenue line. Dell +9.89% to $484.56 and Hewlett Packard Enterprise +8.12% to $58.80 followed, helped by Goldman's Katherine Murphy reiterating Buy on Dell with the target raised to $510, roughly 16% above Tuesday's close, "expecting strong orders and a growing backlog even as supply constraints limit shipments," and raising NetApp to $210 while cutting HPE's to $75 from $79. Seagate +7.05%, Western Digital +3.69%, Micron +4.93%, Arista +6.39%, Teradyne +6.12%, Lam +4.72%, Applied Materials +4.33%, KLA +3.89%, Nvidia +3.03%. The funding leg is the point: Microsoft −2.26%, Meta −3.38%, Amazon −1.83%. The market did not add risk — it moved risk from the companies that spend on AI to the companies that sell the boxes. That is a margin-transfer trade, and it is the second time in three sessions the tape has paid the supplier rather than the buyer. Forward catalyst: Applied Materials 8/13 AMC (§5), the only large-cap semi-cap read in the window.

2. Cisco beat and fell, and it is the cleanest warning attached to item 1. Cisco closed +2.92% at $123.95 in the regular session on the hardware read-across, then fell about 3% after the close: fourth-quarter adjusted gross margin 66.3% against a 66.0% StreetAccount consensus, and Bloomberg's headline was "Cisco's AI Outlook Disappoints Investors Seeking Bigger Payoff." Coherent lost almost 3% post-close with non-GAAP gross margin 40.2% against 40.0%, despite first-quarter guidance above estimates. Cerebras fell 14% on second-quarter revenue of $180m against a $194m LSEG consensus. Three AI-adjacent names beat or met on the headline and all three were sold on the composition. Set that against Nebius +34.14% to $259.20 on a 514% jump in AI cloud sales (Bloomberg) and the dispersion inside the theme is now enormous. Forward catalyst: the Thursday open, which marks Cisco, Coherent and Cerebras against a SOX that closed 2.49% higher.

3. The whole tape faded from the open, and the Dow finished red. The S&P opened up 0.5%, the Nasdaq up 0.9% and the Dow up 151 points (CNBC, 9:30 a.m.); the Dow's high was 53,969.36, 177.51 points above Tuesday, and it closed 21.58 lower. The Nasdaq 100 opened 29,875.18 and closed 29,742.60. The SOX opened 12,518.9 and closed 12,399.4. Four indices, four opens within a fraction of a percent of the day's high, four closes materially below. The mechanism is straightforward: the CPI reaction was a gap, not a trend, and there was no incremental buyer after 10 a.m. For a long/short book the actionable form of this is that intraday momentum has been negative for three consecutive sessions while the index level has gone nowhere — a distribution pattern rather than an accumulation one. Forward catalyst: PPI 8:30 Thursday (§7); whether the next gap holds past the first hour.

4. The rate-sensitive consumer would not participate on the day the hike got priced out. Home Depot −3.12% to $343.43 was the second-largest Dow drag; Lowe's −2.39%, DR Horton −3.30%, Lennar −2.68%, PulteGroup −2.46%, Builders FirstSource −3.58%, Tractor Supply +3.23% the lone exception. This happened on a session in which the 3-year yield fell 2 bp and Investing.com's September hike probability fell 9.7 points. Two days ago this same complex rallied 2.3–4.3% on a 4 bp fall in the 3-year. A group that trades duration on Tuesday and ignores it on Wednesday is trading something else — most plausibly the 8/18 Home Depot and 8/19 Lowe's prints (§5), into which positioning is being cut rather than added. Forward catalyst: Home Depot 8/18 BMO, Lowe's and Target 8/19 BMO.

5. Volatility collapsed into the next print, and Bloomberg named the behaviour. VIX 14.55, −4.78%, the lowest close of the month, with a session range of 14.39–15.42 and the low set after the CPI reaction had been fully absorbed. Bloomberg's markets headline for the session: "Investors Abandon Hedges in Order to Chase Soaring Stocks." The rates side agrees rather than disagrees this time — the MOVE index at 72.09 is the lowest reading in this report's recent range, putting MOVE/VIX at 4.96× (§9c). Both vol markets sold off on the same day, into a PPI print and a retail-sales print 48 hours apart. At 14.55, one-month implied prices a daily S&P move of about 0.92%; realised over the last three sessions is −0.32%, +0.26% and −0.06%, so the short-vol carry is still working, which is exactly why it is crowded. Forward catalyst: PPI and claims 8/13 8:30; retail sales and Michigan preliminary 8/14 (§7).

6. Crude gave back a 1.3% morning gain on an IEA demand cut, and the airlines still fell. WTI was $84.30 pre-open on the first fatal Red Sea shipping attack in over a year — Houthi rebels killed six on a cargo ship in the Bab el-Mandeb — and a U.S. missile strike on a container ship allegedly running the blockade of Iranian ports in the Gulf of Oman. It settled $83.27, up seven cents. The reversal agent was the IEA, which cut its 2026 global demand forecast to −1.6m barrels a day, 510,000 b/d worse than its July estimate, explicitly attributing the deterioration to the Hormuz closure. The second-order tell inverts Tuesday's: United −0.93%, Delta −0.53% and Southwest −1.19% all fell on a day crude was flat, having all risen on a day crude gained 1.30%. The airlines are not trading fuel in either direction; they are trading the reopening headline, and Wednesday's headline was demand destruction rather than diplomacy. Forward catalyst: any SNSC or U.S. statement on Hormuz; the next IEA/OPEC monthly.

7. Managed care rallied as a bloc with no single attributable catalyst. Molina +5.92% to $206.06, Humana +4.29% to $388.83, Centene +3.38% to $67.07, Elevance +2.23% to $399.06, UnitedHealth +0.85%. Five names, one direction, a 0.85%-to-5.92% spread that scales with beta rather than with news, and no company-specific item in CNBC's movers lists or the Bloomberg and WSJ boards. Set against it, the drug-distribution complex broke the other way: Cencora −5.91% to $314.14 — attributed by Investing.com to confirmation that certain Walgreens volume previously handled outside its prime-vendor agreement began shifting to rival distributors on 1 July — and McKesson −2.64%, Cardinal Health −2.54%. Payers up, distributors down, on the same day, is a margin-transfer inside healthcare and worth a pair rather than a directional view. Forward catalyst: the Cencora guidance revision, if one comes; any ACA-market or rate-notice headline.

8. Wendy's jumped on a take-private report, and it is the first LBO headline of the month. The stock rose more than 13% intraday and held above +11% after the Financial Times reported that Nelson Peltz's Trian Fund Management is preparing a bid with co-investors Flynn Group and BlueFive Capital; no timeline is set and the bid is not certain, and Wendy's said it would review any proposal consistent with its fiduciary duties (CNBC). (Not an S&P 500 constituent on the board used for this report.) The reason it belongs in a cross-asset note is the financing environment rather than the name: an HY credit spread at 272 bp, nine basis points tighter year to date, is the condition that makes a leveraged restaurant buyout printable (§9). Forward catalyst: whether a second sponsor situation surfaces this week; HY primary supply.

Eight green, three red, and a 2.67-point dispersion — the narrowest of the week and the first session since Friday with more than four groups higher. Technology +1.38% and Industrials +1.37% led by a nose, and the composition is the story rather than the ranking. Technology's leadership is entirely hardware: Super Micro +19.02%, Dell +9.89%, Hewlett Packard Enterprise +8.12%, Seagate +7.05%, Arista +6.39%, Teradyne +6.12%, Oracle +5.37%, Akamai +5.24%, Corning +5.18%, Micron +4.93%, Lam +4.72%, Applied Materials +4.33%, KLA +3.89%, Western Digital +3.69%, Intel +3.32%, Nvidia +3.03%, Cisco +2.92%, against Microsoft −2.26% and Gartner −4.17%. Seventeen of the top twenty S&P 500 performers were hardware, semis or semi-cap. Industrials +1.37% is the group that will not reconcile arithmetically (below) but its internals are clean: Expeditors +4.10%, JB Hunt +3.99%, GE Vernova +2.77%, Honeywell +2.27% reversing Tuesday's −5.27%, Johnson Controls −1.27% the notable laggard. Consumer Cyclical −1.29% was the worst group on a soft-inflation day, which is the session's most counterintuitive number: Uber −4.05%, Lululemon −3.77%, eBay −3.50%, Take-Two −2.99%, Home Depot −3.12%, Lowe's −2.39%, DR Horton −3.30%, Lennar −2.68%, Deckers −2.81%, Airbnb −2.62%, against Darden +4.01%, Yum! Brands +3.86% and Hilton +2.50%. Restaurants and hotels up, retail and housing down — the consumer split by ticket size, not by rate sensitivity. Communication Services −0.88% has a new driver: Meta −3.38% to $578.85 and Charter −4.78% to $150.15 did the damage while Alphabet A closed −0.07% at $343.57, effectively flat after four losing sessions in five. The group is now −2.02% year to date and the worst YTD performer on the board alongside Consumer Cyclical at −2.37%. Real Estate +0.92% was the third-best group on a session the 10-year fell only 2 bp — Equinix +3.58%, Digital Realty +3.28%, Ventas +2.36% and Welltower +1.88% recovering Tuesday's unexplained 5.43% and 3.84% declines, which retrospectively marks that pair move as a flow event rather than a fundamental one (§2 item 8 of the 8/11 edition). Energy +0.10% on a flat crude settle, with Marathon Petroleum +3.52% and Valero +1.94% again outperforming ConocoPhillips +1.11%, Phillips 66 +0.54%, Chevron −0.04%, Diamondback −0.43%, Occidental −0.86% and APA −1.58% — the refiner-over-producer split is now six sessions old. Source: Finviz Groups via rendered page (?g=sector&v=140&o=name); Finviz buckets are not official GICS/S&P sector indices — they screen all U.S.-listed names, which is why Alphabet and Meta sit in Communication Services and Amazon and Home Depot in Consumer Cyclical.

Reconciliation. All eleven groups were checked against Tuesday's published YTD compounded by Wednesday's 1-day move. Ten reconcile to within 0.02 point: Basic Materials 1.1691 × 0.9967 = +16.52% vs 16.52% (exact); Communication Services 0.9885 × 0.9912 = −2.02% vs −2.02% (exact); Consumer Cyclical 0.9891 × 0.9871 = −2.37% vs −2.37% (exact); Consumer Defensive 1.0754 × 1.0037 = +7.94% vs 7.94% (exact); Energy 1.3394 × 1.0010 = +34.07% vs 34.07% (exact); Technology 1.2362 × 1.0138 = +25.33% vs 25.33% (exact); Financial 1.0823 × 1.0032 = +8.58% vs 8.57%; Healthcare 1.0833 × 1.0018 = +8.52% vs 8.54%; Real Estate 1.0878 × 1.0092 = +9.78% vs 9.79%; Utilities 1.0194 × 1.0043 = +2.38% vs 2.37%. The exception is Industrials, and it is large: 1.1599 × 1.0137 = +17.58% against the +16.22% shown, a 1.36-point gap. A one-day move of +1.37% that adds only 0.23 points to the year-to-date is arithmetically impossible in a fixed basket, so the Finviz Industrials constituent set changed between the two pulls. The 1-day and 1-week figures are still usable; the Industrials YTD is not comparable to the prior edition's and is flagged rather than reconciled. Six groups reconcile exactly; maximum deviation excluding Industrials is 0.02 point (Healthcare).

Up

Down

Times are ET. Every day page from Thursday 8/13 through Friday 8/21 was re-pulled from Earnings Whispers this session (/1 = before open, /2 = after close) and screened against the Investing.com U.S.-500 constituent board. The 8/10–8/12 rosters are carried from prior verified pulls with reactions added from this session's boards. Re-verify times and membership against company IR before trading any date.

Mon 8/10 — completed. BMO: Berkshire Hathaway B (BRK.B) 8:00 — closed +1.46% on the day and has now fallen three consecutive sessions to $510.00, −3.7% from that close. AMC: Simon Property Group (SPG) 4:05.

Tue 8/11 — completed. BMO: Cardinal Health (CAH) 6:45 — closed −2.54% at $234.16 on Wednesday, giving back Tuesday's gain; Amentum (AMTM) 8:00. AMC: Lumentum (LITE) 4:00; Super Micro Computer (SMCI) 4:05 — +19.02% to $37.61, the best performer in the index (§4).

Wed 8/12 — completed. BMO: Amcor (AMCR) 6:00; Trimble (TRMB) 6:55 — closed −2.85% at $56.33. AMC: Cisco (CSCO) 4:05 — closed +2.92% at $123.95 in the regular session, then fell about 3% after the bell on 66.3% adjusted gross margin against a 66.0% StreetAccount consensus; Coherent (COHR) 4:05 — lost almost 3% in extended trading on 40.2% non-GAAP gross margin against 40.0%, despite above-consensus first-quarter guidance.

Thu 8/13. BMO: Tapestry (TPR) 6:45 — the stock has fallen 6.7% over the two sessions into the print. AMC: Applied Materials (AMAT) 4:00 — closed +4.33% at $548.39 the day before.

Fri 8/14. Neither page lists an S&P 500 reporter — the before-open page is entirely micro-cap, biotech and broadcasting, and the after-close page is micro-cap; none of the reviewed names is carried by the constituent board.

Mon 8/17. Neither page lists an S&P 500 reporter (ten names screened across both buckets).

Tue 8/18. BMO: Home Depot (HD) 6:00. AMC: Keysight Technologies (KEYS) 4:05, Jack Henry & Associates (JKHY) 4:15.

Wed 8/19. BMO: Lowe's (LOW) 6:00, Estée Lauder (EL) 6:00, Target (TGT) 6:30, Analog Devices (ADI) 7:00, TJX Companies (TJX) 7:30. AMC: Nordson (NDSN) 4:30.

Thu 8/20. BMO: Deere & Company (DE) 6:20, Walmart (WMT) 7:00. AMC: Ross Stores (ROST) 4:00.

Fri 8/21. Neither page lists an S&P 500 reporter (four names screened before the open; the after-close page returned "NONE").

Changes vs. the prior calendar (8/11 report):

The shape and the diagnostic: a front-led bull steepener of very small amplitude, and it is the smallest reaction to an in-line CPI the curve could plausibly have produced. The tenors that fell most are the 1.5-month, the 2-month and the 1-year, all −3 bp; the tenors that fell least are the 5-year (−1), the 20-year (−1) and the 30-year (0). Read that as two separate segments doing two separate things. The policy-path segment — everything inside two years — rallied 2 to 3 bp, and that is the mechanical consequence of Investing.com's September hike probability falling 9.7 points and CME's September hold rising 8.3 points (§8). The term-premium segment did essentially nothing: 2s30s widened 2 bp to 104 bp, and the long bond finished unchanged for the first time in four sessions. An inflation print that lands on consensus removes near-term policy risk and tells you nothing about the long-run supply-demand balance for duration, and the curve priced exactly that distinction. What it did not do is validate a rally: the S&P rose 0.26% and the 10-year fell 2 bp, which is a 1-to-8 ratio against the 7 bp/0.32% relationship of Monday.

The week is the more important frame, and it points the other way. Over the five sessions since 5 August the front of the curve has fallen — 2-month −4 bp, 1-year −3 bp, 3-month −2 bp — while everything from five years out has risen 5 to 7 bp. 3M10Y has widened 7 bp in a week and 2s30s 5 bp. That is a textbook bear steepener at the back sitting on top of a bull flattener at the front, and the composite reads as: the market has deferred the next hike and simultaneously demanded more compensation for owning duration. An in-line CPI did not reverse a single basis point of the week's term-premium build. For a real-money book that is the trade-relevant fact — the front end is doing what the Fed card says, and the long end has stopped listening to the Fed card altogether.

Bills separately, because they are a funding signal rather than a policy one (§9b). The 3-month par yield fell 2 bp to 3.87% after two sessions unchanged, but the 2-month fell 3 bp and the 1-month only 1 bp, so the very front of the bill curve steepened between one and two months and flattened between two and three. WSJ's real-time 3-month bill was unchanged at 3.802% at 8:56 PM ET; the gap to the official par 3.87% is the usual bank-discount-versus-coupon-equivalent basis and is not a level dispute. Vendor cross-check on the coupon curve: Bloomberg's board marks the U.S. 10-year at 4.67%, −2 bp; WSJ real-time at 8:56 PM ET has 4.689% (−1.2 bp), 30-year 5.249% (−1.2 bp), 3-year 4.265% (−2.6 bp) and 2-year 4.199% (−0.8 bp). WSJ's marks are taken nearly five hours after the official 3 p.m. par fixing and are directionally identical; the 2-year discrepancy of 1.2 bp is a timing artefact, not a level dispute.

Current week (Aug 10–14) — released

Current week — remaining

Next week (Aug 17–21)

Current target range: 3.50–3.75% (WSJ Consumer Rates table, unchanged for the period shown; IORB 3.65%). The market is pricing hikes, not cuts: the probability of a target range below 3.50–3.75% is 0.0% at every 2026 meeting on both vendors.

The headline — CME FedWatch four-column table for the 16 September meeting

Provenance of every column, stated. CME published a complete numeric four-column table under the next-meeting chart on this run, so all four columns are read live from CME's own table rather than reproduced from a prior edition; the reference dates in the column headers are CME's. The footnote reads "Data as of 12 Aug 2026 07:48:00 CT" — 8:48 a.m. ET, eighteen minutes after the CPI release. That is CME's own snapshot time and it is not a live 9 p.m. read, so the NOW column captures the immediate CPI reaction and not the full session; it is disclosed here rather than presented as a close.

The cross-check — Investing.com Fed Rate Monitor, and the gap reconciled

Investing.com's card for 16 September, updated 8:35 p.m. EDT on 12 August, shows hold 62.3% (previous day 52.6%, previous week 46.3%) and +25 bp 37.7% (47.4%, 53.7%), on a September fed funds contract price of 96.325.

The gap is 2.4 percentage points on the hold, and it is a timestamp gap rather than a methodology gap. CME's snapshot is stamped 8:48 a.m. ET; Investing.com's is 8:35 p.m. ET, eleven and three-quarter hours later, and in between the September contract continued to rally as the session digested the print. Quantify it: the 16 September meeting falls on day 16 of a 30-day contract month, so a full 25 bp hike moves the September average effective rate by roughly 25 × (14/30) = 11.7 bp, which means one basis point of ZQU6 is worth about 8.6 points of headline probability. A 2.4-point gap is therefore about 0.28 bp of contract price — well inside a normal overnight drift and far too small to indicate any disagreement between the two calculations. Independent corroboration: CNBC quoted CME FedWatch during the session at "a 73% chance the Fed will raise rates in December" and October hike odds "more than 53%"; Investing.com's cumulative figures for the same two meetings are 72.5% and 53.4%, within a point of both.

(a) Current-year meeting distributions — current / [prev day] / [prev week]

Modal ranges in bold. Contract prices: ZQU6 96.325, ZQV6 96.265, ZQZ6 96.130.

The one-day and one-week changes, and the multi-day momentum. In one session the September hike probability fell 9.7 points (47.4% → 37.7%), the October cumulative 8.6 points (62.0% → 53.4%) and the December cumulative 5.2 points (77.7% → 72.5%). Over one week the same three fell 16.0, 12.1 and 7.5 points. The decay is monotonic in tenor — the further out the meeting, the less of the hike was removed — which is the signature of deferral rather than cancellation. The momentum is now five sessions old and one-directional: CME's own columns put the September hold at 30.4% a month ago, 45.6% a week ago, 51.6% a day ago and 59.9% on its post-CPI snapshot, so the hold has doubled in a month and become modal by a comfortable margin for the first time in the report's window. The other thing CME's month column shows is that a +50 bp September, at 18.4% on 10 July, is now 0.0% — the hawkish tail has been fully extinguished, and with the dovish tail pinned at zero the September distribution is now a clean two-outcome bet for the first time this cycle.

(b) Next-year meeting path

The contract prices are what draw the terminal rate, and they moved today. The lowest price on the strip is 95.950 at the 28 July 2027 meeting, an implied 4.050% — down from 4.075% on Tuesday and 4.100% on Monday, so the terminal has fallen 5 bp in two sessions. The modal range is 3.75–4.00% at every single 2027 meeting, and the modal probability decays monotonically from 41.7% in January to 32.5% in December — the distribution flattens rather than shifts, which is the market saying it knows roughly where the peak is and not when. The first non-zero probability of a target range below 3.50–3.75% anywhere on the curve appears at the 27 October 2027 meeting, at 1.4%, rising to 3.6% by December 2027. Fourteen months of futures with no easing priced at all.

(c) Year-end probability ladders

Year-end 2026 (9 December 2026 meeting), relative to the current 3.50–3.75%:

Year-end 2027 (8 December 2027 meeting), relative to the current 3.50–3.75%:

The week-over-week arc, and it is a divergence. Over five sessions the market removed 16.0 points from a September hike and 7.5 points from the December 2026 cumulative — and over the same five sessions the cumulative probability of sitting above 3.50–3.75% at the end of 2027 rose from 73.9% to 78.3%, while the cumulative probability of sitting below it fell from 6.2% to 3.6%. The front end priced hikes out and the back end priced them in. That is deferral, not cancellation, and it is the single most important thing in this section: an in-line CPI bought the Fed time, and the futures strip charged the market for the time by raising the odds that the eventual destination is higher than 3.75% and shaving 2.6 points off the already-negligible chance of any easing inside sixteen months.

Rounding, stated transparently. Investing.com's cards do not always sum to exactly 100.0%: the September card sums to 100.0, October to 99.9, December 2026 to 100.0, January 2027 to 100.0, March 2027 to 100.1, April 2027 to 100.0, June and July 2027 to 100.1 each, October 2027 to 99.8 and December 2027 to 99.9. The CME September table sums to 100.0. Cumulative figures above are computed by summing the vendor's published cells without re-normalising, so a cumulative may carry the same ±0.1 to ±0.2 rounding as its parent card.

(a) IG and HY credit spreads

As-of date, stated rather than implied: the ICE BofA option-adjusted spread series below are FRED's 11 August prints. FRED publishes these with a one-business-day lag, so there is no 12 August index credit spread available at the time of writing; the same-day direction is cross-checked against Bloomberg and WSJ coverage below.

CDX retrieval note — the six-step ladder was worked again and the index level remains publicly unobtainable. (1) Bloomberg in Chrome: /markets and /markets/rates-bonds were both rendered; the latter publishes the Bloomberg Fixed Income Indices and a full sovereign 10-year grid but no CDX quote, and the Bloomberg quote page for the standard CDX IG identifier returns "no page to display for the security… it may only be available on the terminal." (2) WSJ Market Data Bonds & Rates: Treasurys, money rates and a consumer-rates table — no CDX. (3) ICE / S&P Global / Cbonds index pages: the CDX.NA.IG 5Y page was rendered directly and displays the level as *** bps with the previous value masked to *** on 10/08/2026 — entitlement-gated. (4) FT Markets Data / Reuters credit wraps: no CDX level in the reviewed 12 August material. (5) TradingView / Barchart / CME CDS index product pages: symbols carried, prices masked. (6) Cash-market proxy, labelled as such: the Bloomberg U.S. Aggregate index rose 1.65 points with a 0.42 month-to-date return on a session in which coupon Treasury yields fell 1–2 bp — a duration-only explanation covers most but not all of that, leaving a small residual tightening. All six steps are named so the gap is auditable rather than asserted. Quoting convention reminder: 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.

What the numbers say, and the tail is the story again. The IG credit spread widened 1 bp to 79 bp — the first move in seven prints, and it takes the series back to exactly where it started the year. The HY credit spread widened 2 bp to 272 bp, still 9 bp inside January and within 2 bp of its 2026 tight. The CCC credit spread widened 9 bp to 1,023 bp, four and a half times the HY move, and the CCC-minus-HY differential widened 7 bp to 751 bp against 604 bp in January and 746 bp a week ago. That differential has now widened in three of the last four prints and is 8 bp wider than the 743 bp low it set on 7 August, which reverses the narrowing trend the prior four editions tracked and moves §12's protection trade away from its stated cut trigger rather than towards it. The mechanism is worth naming precisely: an aggregate IG index at 79 bp is a statement about 500 issuers, most of which did not move; a CCC index widening 9 bp on a session when the S&P rose and the SOX gained 2.49% is a statement about the marginal borrower's access to funding, and those two statements are diverging. Same-day direction cross-check: neither Bloomberg's 12 August markets coverage nor WSJ's bonds page reported a credit-spread move of any size on the session, which is consistent with a 1–2 bp index drift and with the tail move being invisible at the index level.

(b) Money-market and funding plumbing

Basis, stated: the NY Fed publishes reference rates at approximately 8:00 a.m. ET for the prior business day, so the SOFR and EFFR figures below are 11 August effective rates published on 12 August. ON RRP and IORB are same-day (12 August); reserve balances are the weekly average for the week ended 5 August.

The distribution is doing something the level is not, and the buffer has now formally run out. The secured mean rose 1 bp to 3.64%, the 75th percentile rose 2 bp to 3.70% and the 99th 2 bp to 3.73%, while the 1st percentile held at 3.58% — so the 99th-minus-1st span widened from 13 bp to 15 bp, the widest in the report's window, for a second consecutive session of widening. A rising mean with a rising top tail and a flat floor is dispersion between cash-rich and cash-poor counterparties, and it is now a two-session pattern rather than a one-session observation. The context that makes 2 bp worth a paragraph: ON RRP printed $0.725bn, a new record low, four sessions after breaking below $1bn for the first time ever, and reserve balances are $149.4bn below the July peak. The facility is empty, the reserve buffer is shrinking, and the secured market is beginning to price counterparties differently. None of this is stress — SOFR is one basis point through IORB and repo volumes over $1.2tn a day are not distressed — but the sequence is the right one, and the September quarter-end is the test. The one scheduled release that speaks directly to this is the NY Fed's Reserve Demand Elasticity print on 20 August (§7). Note also that EFFR volume at $109bn is near the bottom of the reviewed window while the unsecured percentiles did not move at all: the unsecured market is clearing less because there is nothing to arbitrage, not because it is impaired.

(c) Rates volatility and swap spreads

Both vol markets sold off together, and that is different from the last two sessions. For a fortnight this report has had to caveat the MOVE as a stale series pairing a current numerator with an old denominator. On this run the series moved: 72.09 against a 77.92 implied prior, −7.48% on the card, and it is the lowest rates-vol reading in the report's window. The VIX fell 4.78% to 14.55, its lowest close of the month. So the ratio at 4.96× is flat while both legs compressed — rates vol and equity vol were sold on the same session, into a producer-price print eleven hours later and a retail-sales print thirty-six hours later. Bloomberg's own headline for the day was "Investors Abandon Hedges in Order to Chase Soaring Stocks," which is the behavioural description of the same fact. At 14.55 one-month S&P implied prices a daily move of roughly 0.92%, against realised of −0.32%, +0.26% and −0.06% over the last three sessions — implied is still comfortably above realised, so the short-vol carry is working, which is precisely why the position is crowded. The honest gap in this table remains swap spreads, and it matters this week: §6 shows the 30-year unchanged on an in-line CPI and 7 bp higher on the week, which is a term-premium move that would show up first in the 30-year swap spread, and no public source publishes it.

(d) Issuance, leveraged loans and private credit

Quote basis — read this before the table. All pairs are spot in the market convention shown; a positive move on a USD/XXX pair means the dollar strengthened, and on EUR/USD, GBP/USD, AUD/USD and NZD/USD it means the dollar weakened. The TradingEconomics pull landed at approximately 09:04 on the vendor's clock, i.e. 21:04 ET on 12 August, after that vendor's daily boundary had rolled: most rows carry an "Aug/13" stamp and the vendor's own %Chg column prints between −0.30% and +0.02% for the whole board. That column is therefore NOT reproduced. The "24h" column below is computed by this report against the same vendor's levels published in the 11 August edition, taken at 18:20 ET — an interval of roughly 26.7 hours rather than exactly 24, which is disclosed here and worked in Data Notes. YTD figures are the vendor's own column.

The take — the dollar rose on a soft print, and that is the whole FX story. Run the inputs: the 1-year Treasury yield fell 3 bp, the 2-year 2 bp, Investing.com's September hike probability fell 9.7 points and the 2027 terminal contract price implies a rate 2.5 bp lower than Tuesday's (§8). Every one of those is dollar-negative, and the dollar index closed 0.13% higher over 24 hours — the largest move in four sessions, with the wrong sign. The prior edition described an FX market that had "stopped taking a view before an 8:30 print" and was flat; this session shows what flat books do after the event, which is buy the currency because the near-term policy risk has been removed rather than sell it because the rate has fallen. The mechanism is the terminal rate, not the front end. §8 shows the market deferring hikes at the front while raising the cumulative odds of sitting above 3.50–3.75% at end-2027 from 73.9% to 78.3% over the week. A currency trades the destination, not the timing, and the destination got firmer. That is the single cleanest reconciliation available between two markets that appear to disagree, and it is why the dollar's response was not a mistake.

The contrarian cross is the won, and it is the exact mirror of Tuesday. USD/KRW rose 0.08% — the won weakened — on a session when the Kospi gained 3.68% to 6,579.04 on Samsung Electronics +7% and SK Hynix +6.6%, and Bloomberg reported Korean stocks up 22% in ten days. On Tuesday the won strengthened 0.26% on a day the Kospi reversed to close +0.73%. Take the sequence over four sessions: the won strengthens when the large-cap index reverses on genuine foreign semiconductor flow, and weakens when it melts up. A 22% ten-day equity rally that is not pulling the currency with it is a domestically-funded rally, and the leverage sits in Seoul rather than offshore. That distinction is available in real time in a market that trades around the clock, and it is the highest-quality tell on this board.

The franc is now a three-session pattern and it has survived every alternative explanation. USD/CHF rose 0.24%, the largest of the three moves, on a day with no oil shock, no equity decline, a VIX at a monthly low and an inflation print that came in on consensus. Against it, gold settled 0.90% higher for a sixth consecutive gain (§11). The market is buying the non-yielding real asset and selling the low-yielding haven currency, which is the inflation-hedge configuration rather than the risk-off configuration — and it has now held for four sessions through a fading crude shock, a soft CPI and a 4.78% VIX decline. Any one of those should have broken it if the driver were positioning; none did. §12 keeps the expression, at reduced size, on the entry rather than on the thesis.

And the one that should have moved and did not: USD/CAD at +0.15% on a crude settle of plus seven cents. The magnitude is fine; the sign is wrong for a petro-currency on a flat barrel and it reverses Tuesday's correctly-signed −0.10% on a 1.30% crude gain. Two sessions, two opposite signs, on a commodity that has gone nowhere — the loonie is trading the U.S. rate differential (2-year −2 bp) rather than the barrel, and anyone still using CAD as a crude proxy is running an unhedged rates position.

Basis, stated before the table. The settlement column is the 12 August front-month settle, derived where necessary from two independent vendors (CNBC contract "Prev Close" during the 13 August session, cross-checked against TradingEconomics' and WSJ's implied prior — the derivation is worked in Data Notes). The YTD column is TradingEconomics' own spot-basis series and is NOT on the same basis as the settlement column; the two are not arithmetically reconcilable and the vendor column is reproduced with attribution rather than recomputed. Contract months are stated in every row.

The take — crude round-tripped a war headline into a demand cut, and that is the most important commodity fact of the session. WTI was $84.30 pre-open, up 1.3%, on two genuinely escalatory events: Iran-backed Houthi rebels killed six people in an attack on a cargo ship in the Bab el-Mandeb, the first reported fatalities from Red Sea shipping attacks in more than a year, and hours later U.S. forces fired missiles at a container ship allegedly attempting to break Washington's blockade of Iranian ports in the Gulf of Oman. It settled $83.27, up seven cents. The agent of the reversal was the IEA's monthly report, which cut its 2026 global demand forecast to −1.6m barrels a day, 510,000 b/d worse than July's estimate, explicitly attributing the deterioration to the Hormuz closure and noting that high fuel prices will continue to weigh on consumption before demand returns to growth in the fourth quarter. The market has now been handed the second-order consequence of its own first-order shock and has decided it is the larger number. For positioning that is a regime marker: for eight weeks the barrel has traded every Hormuz headline in one direction, and this is the first session in which a genuinely escalatory pair of events could not hold a 1.3% gain.

The positioning read, and the crack is still where the money is. Distillate and refining margin remain the cleanest expression: Marathon Petroleum +3.52% and Valero +1.94% against ConocoPhillips +1.11%, Phillips 66 +0.54%, Chevron −0.04%, Diamondback −0.43%, Occidental −0.86% and APA −1.58% — six consecutive sessions of refiners over producers, on a barrel that has gained 0.16% across the last two. The mechanism is that a logistics disruption widens the crack while doing nothing for the wellhead, and gasoline at +82.75% year to date on the vendor's own column against crude at +42.99% is the same fact at the product level. The airlines confirm the direction rather than the level: United −0.93%, Delta −0.53%, Southwest −1.19% on a flat barrel, having risen on a barrel up 1.30% the day before (§2 item 6).

Gold's sixth consecutive gain is the number that does not fit, and it is the one to keep. An in-line CPI that cut the September hike probability by 9.7 points should, mechanically, reduce the appeal of a non-yielding asset less than it reduces the appeal of the dollar — and instead the metal settled +0.90% at $4,467.50 while the dollar index rose 0.13% (§10). Both cannot be a rates trade. The reconciliation is that gold is being bought as an inflation hedge and a reserve asset rather than as a rate-differential trade, which is exactly the configuration §10 identifies in the gold-versus-franc pair. Note the vendor basis carefully before trading it: the Comex December futures settle is $4,467.50, TradingEconomics' spot is $4,433.87 and Bloomberg's Asia-hours board ran $4,489.50 to $4,499.00 across the evening. The $33.63 futures-over-spot gap is the December contract's carry, and the Bloomberg readings are live Asian-session prints rather than settlements — all three are correct on their own basis and none should be compared to another without adjusting. Silver's +1.30% against copper's −0.20% is the other tell: the precious side of the complex is bid and the industrial side is not, which is a monetary-demand signal rather than a growth one.

Desk-style ideas for institutional investors. Each carries an explicit catalyst and an explicit invalidation. These are not personalized investment advice; verify independently and size to your own mandate before acting.

1. The rates trade — close the back-versus-front spread, reverse it, and be honest about why

Mark on the prior edition's rates idea, first. The book was long ZQZ7 against short ZQU6, DV01-matched one-for-one, at two-thirds size, from 96.000 / 96.315. It gained 0.5 bp on Wednesday — ZQZ7 rose from 95.975 to 95.995 (+2.0 bp) while ZQU6 rose from 96.310 to 96.325 (+1.5 bp) — and it is now 1.5 bp offside since inception, improved from 2.0 bp. Neither stated invalidation triggered: July core CPI printed +0.2% m/m against the +0.4% trigger, and ZQN7 is 95.950 against the 95.850 trigger. So the trade is alive on its own terms and it is still wrong, and that is the more useful admission. The structure is long the back and short the front — it profits when the market prices near-term hikes and later cuts. §8 documents the market doing precisely the opposite over five sessions: 16.0 points removed from a September hike while the cumulative probability of sitting above 3.50–3.75% at end-2027 rose from 73.9% to 78.3%. A position fighting a five-session, one-directional repricing on both legs simultaneously does not deserve a sixth session on a 0.5 bp win.

Modal path, base case and the tails — this is what the replacement expresses. Modal path: hold on 16 September (Investing.com 62.3%, CME 59.9%), hold on 28 October at 46.5% modal with the cumulative hike at 53.4%, one 25 bp hike delivered by 9 December with 3.75–4.00% modal at 45.0% and the cumulative at 72.5%; terminal 4.050% drawn by the 28 July 2027 contract at 95.950; year-end 2027 modal 3.75–4.00% at 32.5%, with the first non-zero easing probability anywhere on the strip at 1.4% in October 2027. Base case: in-line inflation, claims under 200,000 for a third week, a crude shock the IEA has just told us is destroying demand rather than creating it, and a committee with the room to wait — hikes deferred to December, not cancelled. Hawkish tail: PPI at or above +0.4% m/m on Thursday, which at 8.6 points of probability per basis point of ZQU6 rebuilds roughly 10–15 points of September hike risk and drags the whole 2027 strip lower. Dovish tail: initial claims above 240,000 alongside a soft PPI, which would produce the first non-zero 2026 cut probability of the year — currently 0.0% at every 2026 meeting for an eighteenth consecutive session.

Expression: close the long-ZQZ7/short-ZQU6 spread and reverse the structure — long ZQV6 (October 2026 fed funds) against short ZQZ7 (December 2027 fed funds), DV01-matched one-for-one, both legs 30-day fed funds futures at $41.67 of DV01 per basis point per contract. Entry spread 27.0 bp (96.265 less 95.995). The long leg owns the deferral the data supports; the short leg is the terminal rate the futures strip has been quietly raising all week. Rationale for the reversal rather than a flatten: ZQV6 at 96.265 implies roughly 53% odds of at least one hike by October, so the long leg is close to symmetric with 5.5 bp of upside to a certain hold and 6.5 bp of downside to a certain September hike; the short leg carries the 2027 hawkish drift that has cost the prior structure money for five sessions and is now on the right side of it. Catalyst: PPI and initial claims 8/13 at 8:30; advance retail sales and Michigan preliminary 8/14 at 8:30 and 10:00 (§7). Invalidation: July PPI at or above +0.4% m/m; or the spread through 23.0 bp; or ZQV6 below 96.200. Sizing: small, half the size of the position it replaces, and note the spread's carry is slightly negative — size on roughly 4 bp of maximum adverse spread move.

2. Own short-dated downside, not vol — a put spread beneath Wednesday's low

Honest mark on the prior edition's long-vol idea: it lost, and it lost on the exact invalidation that was written down. The stated invalidation was "the print landing exactly on consensus with a core at +0.2%, in which case the vol is a straight decay loss." Core printed +0.2% and headline +0.1%, both on consensus; the VIX fell 4.78% to 14.55 and the position is a full premium write-off. It was sized as premium-at-risk and it should be closed rather than rolled.

Expression: a one- to two-week S&P 500 put spread struck beneath Wednesday's 7,737.95 low, financed by selling the wing — not an outright straddle and not VIX calls. Thesis: the reason to own downside is no longer the event distribution, which is now known; it is the price of the hedge. VIX 14.55 is the lowest close of the month, MOVE 72.09 the lowest reading in the report's window, and for the first time in a fortnight both vol markets compressed on the same session (§9c) — into a producer-price print eleven hours away and a retail-sales print thirty-six hours away. Bloomberg's own framing was "Investors Abandon Hedges in Order to Chase Soaring Stocks." At 14.55 one-month implied prices a daily move of about 0.92%; the index has opened at or near its high and faded on three consecutive sessions (§2 item 3), which is a distribution pattern rather than an accumulation one. Catalyst: PPI 8/13 8:30; retail sales 8/14 8:30. Invalidation: a PPI at or below +0.1% m/m, which removes the last near-term hawkish hook and leaves the structure to decay; or the S&P closing above 7,766.01, Wednesday's high, on volume. Sizing: small, defined-risk, and prefer the spread to the outright because the skew is not cheap even if the level is.

3. Own the crack, not the barrel — long refiners against short crude

Expression: long a basket of Marathon Petroleum, Valero and Phillips 66 against a short in September WTI, sized so the futures leg matches the basket's crude beta. Mark: six consecutive sessions of the refiners over the producers, and Wednesday was the cleanest yet — MPC +3.52%, VLO +1.94%, PSX +0.54% against COP +1.11%, FANG −0.43%, OXY −0.86%, APA −1.58% and a crude settle of plus seven cents. Thesis, strengthened: the IEA's demand cut is the first authoritative statement that the Hormuz closure is destroying consumption rather than tightening supply, and demand destruction compresses the wellhead while a logistics disruption keeps the crack wide. Gasoline at +82.75% year to date against crude at +42.99% on the same vendor's board is the product-level version of the same fact (§11). Catalyst: the EIA weekly inventory report; the next OPEC monthly. Invalidation: a confirmed Hormuz reopening, which compresses the crack faster than the barrel and loses on both legs; or a distillate build above 3m barrels. Sizing: medium, unchanged — this is the highest-conviction idea in the section.

4. Close the crude-consumer short entirely

Honest mark: the residual position was short the cruise lines against long index, carried after the airline leg was cut on Tuesday. Wednesday: Carnival −0.20%, Royal Caribbean +0.28% against the S&P +0.26% — approximately flat, a rounding error either way. Thesis for closing: the driver has now inverted twice in three sessions. The airlines rose on a barrel up 1.30% and fell on a barrel up 0.08%; the cruise lines did neither. A short whose thesis was fuel cost, in a complex that has stopped trading fuel in either direction, is a residual rather than a position. The stated invalidation of WTI below $78.18 never triggered, so this is a discretionary close on a change in the driver, not a stop. Sizing: to zero.

5. Long gold against the Swiss franc — keep it, keep it small

Expression: long spot gold funded in CHF, notional-matched, at half size since Tuesday. Mark: both legs worked. Gold settled +0.90% at $4,467.50, a sixth consecutive gain; USD/CHF rose 0.24%, the largest of three consecutive sessions of franc weakness (§10, §11). Thesis, and what strengthened: the configuration has now survived every alternative explanation on offer — a fading crude shock, an in-line inflation print, a 4.78% VIX decline and a rising dollar. A haven currency that will not bid across all four of those is not being treated as a haven, and a non-yielding metal that rises on the day the market prices out a hike and buys the dollar is being held for something other than the real rate. That is the inflation-hedge configuration and it is four sessions old. What has not changed is the entry: Bespoke's overbought flag from Tuesday still stands, and the metal is up roughly 9.3% on the month on the vendor's board. Catalyst: PPI 8/13, the direct test of the inflation-hedge leg. Invalidation: the franc strengthening more than 1% against the dollar in a session where gold is flat or lower; or gold breaking below $4,300 spot. Sizing: small, unchanged at half.

6. Protection on the CCC cohort funded in IG — hold, and reset the trigger

Expression: long CCC-exposed credit protection (or short a levered-loan/CCC-heavy vehicle) against long IG cash. Mark: the trade worked and the thesis was vindicated in the direction that matters. The CCC credit spread widened 9 bp to 1,023 bp against IG +1 bp and HY +2 bp; the CCC-minus-HY differential widened 7 bp to 751 bp, is 5 bp wider than a week ago and 8 bp wider than the 743 bp low of 7 August. The narrowing trend the prior five editions tracked has broken. Thesis: the tail borrower is losing funding access on sessions when the index is making highs, which is what a late-cycle credit market looks like from the inside. Catalyst: PPI 8/13; the September quarter-end funding test, with ON RRP now at a record-low $0.725bn and reserves $149.4bn off the July peak (§9b). Invalidation, reset: the differential back through 735 bp, which would restore the narrowing trend and retire the idea; the prior 730 bp cut trigger is replaced. Sizing: small, restored from a quarter to a half of the original — the widening earns size back, but the FRED series lags a business day and the position should not be sized on a number that is 36 hours old.

7. Cut the electrical-equipment-versus-semiconductor pair — it had its worst day

Expression: long a basket of Eaton, Emerson, Hubbell, Rockwell Automation and Johnson Controls against short an equal-dollar SOX exposure. Honest mark: the pair lost approximately 2.85 points in a single session. The basket averaged −0.36% (ETN +0.15%, ROK +0.98%, EMR −0.35%, HUBB −1.29%, JCI −1.27%) against the SOX +2.49%. Over the two sessions since inception the pair is net −0.70 points, because Tuesday's +2.15-point gain has been more than reversed. The stated invalidation — the SOX outperforming the basket by more than 4% over five sessions — has not triggered, at 2.85 points over two. Thesis, under pressure: the argument was that a $500bn financing coalition changes who pays for the build-out rather than who sells the chips, and that the constrained input is switchgear and cooling rather than silicon. Wednesday's tape said the constrained input is memory, storage and servers — Micron +4.93%, Seagate +7.05%, Dell +9.89%, Super Micro +19.02% — and that is a genuine challenge to the thesis rather than noise. Catalyst: Applied Materials 8/13 AMC (§5), the semiconductor leg's direct test. Invalidation, unchanged: the SOX outperforming the basket by more than 4% over five sessions; or Nvidia's five-year CDS re-widening through 82 bp. Sizing: cut from small-to-medium to small, and beta-adjust rather than dollar-match — the basket's beta sits well below the SOX's and dollar-matching has been flattering the loss.

8. New — the consumer split by ticket size, into next week's retail block

Expression: long a basket of Darden, Yum! Brands and Hilton against short a basket of Home Depot, Lowe's and the homebuilders (DR Horton, Lennar, PulteGroup), beta-adjusted rather than dollar-matched. Thesis: Wednesday produced the cleanest consumer dispersion of the month on a session that should have helped the rate-sensitive side and did not. Restaurants and hotels rose 2.0% to 4.0% — DRI +4.01%, YUM +3.86%, HLT +2.50%, CMG +2.00% — while the big-ticket, rate-sensitive complex fell 2.4% to 3.6% — HD −3.12%, LOW −2.39%, DHI −3.30%, LEN −2.68%, PHM −2.46%, BLDR −3.58% — on a day the 3-year yield fell 2 bp and the September hike probability fell 9.7 points. A group that ignores a favourable duration move two days after trading it perfectly (§2 item 4) is de-risking into an event, and the event is the six-reporter consumer block of 8/18–8/20 (§5). Catalyst: advance retail sales 8/14 8:30; Home Depot 8/18 BMO; Lowe's, Target and TJX 8/19 BMO; Walmart and Ross Stores 8/20. Invalidation: a retail-sales control group above +0.6% m/m, which would say the big-ticket weakness is positioning rather than demand; or the short basket outperforming the long basket by more than 3% before 8/18. Sizing: small — this is an event-window trade with a hard stop at the Home Depot print, not a structural short.

Volatility note. VIX 14.55 (−4.78%), range 14.39–15.42, the lowest close of the month; MOVE 72.09 (−7.48% on the vendor's card), the lowest reading in this report's window; MOVE/VIX 4.96×, effectively unchanged because both legs fell together (§9c). That simultaneity is the change worth pricing. For two weeks the asymmetry was cheap equity vol against carried rates vol, which is a hedgeable divergence; on Wednesday both markets sold protection into the same 48-hour window containing a producer-price print and a retail-sales print. At 14.55 one-month S&P implied prices a daily move of about 0.92% against three-session realised of −0.32%, +0.26% and −0.06%, so the carry is still positive and the position remains crowded for a good reason. The specific structural argument for owning downside rather than vol is in idea 2: the index has faded from its open three sessions running while finishing the three-day period roughly unchanged, and a put spread beneath 7,737.95 expresses that pattern without requiring the vol level to rise.

The crowded consensuses, with the numbers attached.

1. "AI hardware is the new leadership." Super Micro +19.02%, Dell +9.89%, HPE +8.12%, Seagate +7.05%, Micron +4.93%, and seventeen of the top twenty S&P 500 performers in hardware, semis or semi-cap. The stress test is that all three after-hours reports from inside the theme were sold: Cisco −3% on a gross margin that beat by 30 bp, Coherent −3% on a margin that beat by 20 bp, Cerebras −14% on a 7% revenue miss. A theme that rallies 2.49% at the index level and sells every actual result inside it is trading on read-across rather than on numbers, and Applied Materials on 8/13 is the first opportunity to find out which.

2. "The Fed is done hiking for now." September hold 62.3% on Investing.com and 59.9% on CME, up from 46.3% and 45.6% a week ago; 2026 cut probability 0.0% at every meeting. The stress test is Thursday's PPI: at 8.6 points of probability per basis point of ZQU6, a +0.4% print rebuilds 10–15 points of September hike risk inside a minute, and the market has already spent five sessions moving one way. The position is not that hikes are cancelled — the strip is pricing more terminal, not less — it is that they are deferred, and deferral is the most fragile of the three possible views.

3. "Vol is a sell." VIX 14.55, the lowest close of the month; MOVE 72.09, the lowest in the window; both fell on the same session. Bloomberg's headline was "Investors Abandon Hedges in Order to Chase Soaring Stocks." The specific fragility is that equity vol and rates vol are now cheap at the same time, which removes the cross-asset hedge that has worked all month — there is no longer a rich leg to fund the cheap one.

4. "Credit is fine." IG 79 bp, exactly where it started the year; HY 272 bp, within 2 bp of its 2026 tight. Underneath, the CCC credit spread widened 9 bp to 1,023 bp, +138 bp YTD, and the CCC-minus-HY differential widened to 751 bp against 604 bp in January, breaking five editions of narrowing. An index that has not moved all year while its tail has widened 147 bp is not evidence of health; it is evidence that the index is the wrong instrument.

5. "Korea is the trade." Bloomberg counts Korean stocks up 22% in ten days; the Kospi rose 3.68% on 12 August and a further 4% at the 13 August open. The tell is the currency: the won weakened 0.08% on the 3.68% session (§10). A double-digit ten-day equity move that will not pull its own currency is domestically funded, and domestically funded rallies unwind faster than foreign-flow ones.

The two-sided geopolitical tape. Escalation was real and specific on Wednesday — the first fatal Red Sea shipping attack in over a year, six killed in the Bab el-Mandeb, and a U.S. missile strike on a container ship in the Gulf of Oman — and crude finished up seven cents. De-escalation has been rumoured all week and the strait remains closed. The market now has evidence that it will not pay up for escalation headlines and has not been paid for de-escalation headlines either, which means the risk is a resolution in either direction rather than a drift. The IEA's 510,000 b/d demand downgrade is the new variable and it cuts against the barrel in both scenarios: a reopening releases supply into weaker demand, and a prolonged closure destroys more of it.

Structural watch items.

Full Source Links, the Overnight / Asia & Europe read-through, and the complete Data Notes & Conflicts appendix are in the companion file US_CrossAsset_Daily_2026-08-12_DataNotes.txt.