U.S. Stock, Fixed Income & Cross-Asset Closing Daily Monday, August 17, 2026 — Full Market Close Report | Data as of: ~6:10 PM ET (Fed-probability cards timestamped 14 Aug 2026 05:02:44 PM CT — see Section 8) Prepared for Institutional Investors. Not Personalized Investment Advice; Verify Independently before Acting. Full Data Notes & Conflicts and Source Links are in the companion file US_CrossAsset_Daily_2026-08-17_DataNotes.txt. | | Tooling constraint, stated up front. The browser tool was unreachable all session, which blocked this report’s normal routes to the Finviz sector table, the live CME and Investing.com probability boards, the vendor FX and metals boards, the official 17 August Treasury curve row and the Earnings Whispers day pages. Every affected field is named where it falls, with the retrieval attempts listed. Nothing has been estimated to fill a gap. |
| 1 · Executive Dashboard | | The tape in one paragraph. Oil took the market down, and the bond market did the carrying. No release rated "Very high" landed in the past twelve hours — Monday's only scheduled U.S. item was the Empire State survey (Medium), and one vendor's calendar showed the docket empty altogether (§7) — and no "Very high" release is due in the next twenty-four hours either, with Tuesday offering housing starts and pending home sales, both Medium. That vacuum is exactly why the session belonged to a single headline. The 17 June memorandum of understanding between Washington and Tehran hit its 60-day deadline with nothing signed; Iran's foreign ministry ruled out an extension and President Trump said he was not interested in one. Brent rose 2.7% to $90.87 and WTI 2.6% to $84.50 (AP, CNBC), and the equity losses, in the Associated Press's phrasing, "solidified in the afternoon when oil prices accelerated upward." The S&P 500 closed 7,745.06, −40.70, −0.52%, the Dow 53,459.78, −272.63, and the Nasdaq Composite 26,644.91, −84.25, −0.32% — an index-level move that badly understates the damage underneath. The Invesco S&P 500 Equal Weight ETF fell 0.58%, more than the cap-weighted index, and decliners beat advancers about 1.5-to-1 on the NYSE (Motley Fool, midday). What kept the Nasdaq from a worse day was two stocks: SK Hynix +6.5%, adding roughly $83bn of market value, and Micron +5.9% back above $1,000. The other side was the AI funding leg, not the AI demand leg. Microsoft was the index's single largest drag at −2.4% after a Morgan Stanley note argued hyperscalers are spending heavily upfront on data-centre builds whose payoff may take years — and the memory names read the same note as bullish. That is the fourth consecutive session in which this report has flagged capital-structure news, not earnings news, as the transmission channel. Nvidia rose only 0.9% on disclosing it will finance up to $105bn in credit and compute for an OpenAI data centre in Pike City, Ohio, a 20-year project whose first 4.25 gigawatts arrive in 2028 — a nine-tenths-of-a-percent move on a $105bn commitment is the tell. Rates did the rest. The 30-year Treasury yield topped 5.31%, its highest in 19 years and the highest since June 2007 (CNBC), and the 10-year climbed to 4.72% from 4.68% late Friday (AP), which has now taken it from 3.97% before the war with Iran began. Three single names carried real information. L3Harris fell 4.6% after Christopher Kubasik stepped down as CEO and chairman following "certain conduct… not consistent with the values of the Company", with the company stating it was unrelated to financial reporting, controls, customer relationships or operational performance. Constellation Brands fell 6.2% after Berkshire Hathaway disclosed it had sold its entire holding, while Alphabet dipped only 0.5% on the same filing showing Berkshire increased its stake — the buy was worth a fifth of what the sell was worth to the tape. And Meta fell 3.3% ahead of Tuesday's opening arguments in the Oakland federal trial in which 29 state attorneys general allege it designed Facebook and Instagram to addict children. The second-order tell is the one to keep: the VIX rose 7.01% to 15.25 on a 0.52% index decline — the first session in weeks in which equity vol has bid at all, and it bid on a geopolitical supply shock rather than on a growth scare, three days after closing at a seven-month low. |
| | Index | Close | Chg | %Chg | Note |
|---|
| S&P 500 | 7,745.06 | −40.70 | −0.52% | 0.92% below Thursday's 7,816.70 record intraday high. The equal-weight RSP fell 0.58%, more than the cap-weighted index (AP; Motley Fool) | | Nasdaq Composite | 26,644.91 | −84.25 | −0.32% | Held up by memory: SK Hynix +6.5%, Micron +5.9%. Microsoft −2.4% was the largest single drag | | Dow Jones Industrials | 53,459.78 | −272.63 | −0.51% | The largest point decline since 30 July. Went red minutes after the open and never recovered | | Nasdaq 100 | No reliable data available at this time | — | — | The Investing.com and Yahoo NDX boards are Chrome-gated and the cached copies returned Friday's close; the Monday futures print of 30,193.50 is a pre-open mark, not a close, and is not published as one | | Russell 2000 | 3,055.03 | −13.38 | −0.44% | An intraday ~2:00 p.m. ET mark, not a verified close — flagged, not smoothed. Friday's fourth straight gain and 52-week high did not extend | | VIX | 15.25 | +1.00 | +7.01% | Same ~2:00 p.m. ET basis. The largest one-day rise in weeks, three sessions after a 29-December low of 14.25 | | PHLX Semiconductor (SOX) | No reliable data available at this time | — | — | The Yahoo quote page served a cached 14 August close stamped 5:15:59 p.m. EDT and Investing.com's board needs Chrome. Not published rather than carried forward | | UST 10Y (real-time) | 4.72% | +4 bp | — | AP, late Monday, against Friday's 4.68% official par. Up from 3.97% before the Iran war | | UST 30Y (real-time) | 5.311% | +4 bp | — | The highest in 19 years, and the highest since June 2007 (CNBC) | | UST 2Y / full par curve | See §6 | — | — | Treasury.gov's 17 August row was not retrievable this session; the official 14 August curve is carried as the anchor and the basis is stated in §6 | | Brent front month | $90.87 | +2.7% | — | Through $90 for the first time this month. The 17 June U.S.–Iran memorandum expired unrenewed (AP) | | WTI front month | $84.50 | +2.6% | — | From Friday's $82.40 settle the arithmetic is +$2.10 / +2.55%; CNBC's +2.6% is quoted as published (§11) | | Gold (Comex front) | $4,477.20 | +$39.90 / +0.90% | — | A ~2:00 p.m. ET mark, not a settle. The Asian-session print was $4,451.30, +$14.00 / +0.32% — both bases shown (§11) | | Silver (spot) | $65.19 | +0.8% | — | Asian-session basis, disclosed rather than dressed as a Comex settle | | DXY | ~99.40–99.50 | −0.14% | — | A third consecutive decline and the lowest since June 2026 on TradingEconomics' series. Asian-hours basis (§10) |
| | --- | | | 2 · Market Hot Spots (ranked by tradability) | | 1. The oil bid is now a rates event, and the rates event is now an equity event. The 17 June memorandum of understanding between Washington and Tehran reached its 60-day deadline on Monday with nothing signed; Iran's foreign ministry ruled out talks to extend it and Trump said he was in no rush. Brent rose 2.7% to $90.87, WTI 2.6% to $84.50, and the AP's own wrap is explicit about the sequence: "Monday's rally for oil prices sent Treasury yields in the bond market higher, which in turn raised the pressure on the economy and prices for all kinds of investments." The 30-year topped 5.31%, a 19-year high, and the 10-year went to 4.72% from 4.68%. The transmission is mechanical and it is not about growth: a higher barrel raises headline inflation, which raises the probability the Fed hikes, which lifts the whole curve, which compresses equity multiples. The 10-year is up 75 bp from 3.97% since the war began. Forward catalyst: Wednesday's July FOMC minutes, and any Hormuz headline — shipping is at a trickle, with three vessels crossing on Sunday against roughly 130 daily before February. | | 2. Memory is now trading as a separate asset class from the hyperscalers, on the same research note. Morgan Stanley published on hyperscaler AI capex — heavy upfront cash, payoff possibly years out, financial strain in between. Microsoft fell 2.4% and was the S&P 500's largest single drag. SK Hynix rose 6.5%, adding about $83bn of market value, and Micron rose 5.9% back above $1,000. Same note, opposite readings, and both are internally consistent: if hyperscalers are committed to spending regardless of the payoff timing, the component supplier is long the spend and the spender is short it. This report has flagged semiconductor dispersion for four consecutive sessions; Monday it stopped being dispersion inside semis and became a clean short-the-buyer, long-the-seller split. Forward catalyst: Nvidia's next print, and any hyperscaler capex guidance revision. | | 3. Nvidia committed up to $105bn and the stock moved 0.9%. The company disclosed it will finance up to $105bn in credit and compute for an OpenAI data centre in Pike City, Ohio, with the first 4.25 gigawatts of a 20-year project arriving in 2028. A vendor-financing commitment of that size moving the equity nine-tenths of a percent means one of two things: the market has already capitalised the arrangement, or it does not believe the cash converts. Set it against Friday's Broadcom episode — a credit action, not an earnings action, taking 5.95% out of the stock — and a pattern is now four sessions old: the AI trade's marginal information is arriving through the funding structure. Forward catalyst: any rating-agency commentary on vendor financing; Broadcom's 2 September print. | | 4. Volatility finally bid, and it bid on the supply shock rather than the demand data. VIX +7.01% to 15.25 on a 0.52% index decline, three sessions after 14.25, the lowest close since 29 December. Friday's setup was the opposite: a consumer print that missed by seven tenths and a VIX that fell 2.60%. The distinction matters for how you hedge. Equity vol would not pay for a weak consumer; it will pay for a closed strait. Forward catalyst: the U.S.–Iran stalemate, and the six consumer prints landing between Tuesday's open and Thursday's close (§5). | | 5. Berkshire's 13F moved two S&P 500 names in opposite directions and the asymmetry is the point. Constellation Brands fell 6.2% after Berkshire disclosed it sold its entire holding in the Modelo and Robert Mondavi distributor; Alphabet dipped only 0.5% on the same filing showing Berkshire had increased its Google stake alongside several homebuilders. A full exit is worth 6.2% and an add is worth −0.5%. That is a market that treats the Berkshire signal as informative on the way out and noise on the way in — which is either a sensible read of position sizing or a straightforward asymmetry to fade. Forward catalyst: the next 13F cycle; any follow-on disclosure on the homebuilder positions into a 5.31% 30-year. | | 6. The dollar is the most crowded trade on the board and it is unwinding into the oil bid. Speculative long USD positioning stands at +$48bn, more than quadruple its level three months ago and the most crowded since 2015, and DXY fell for a third consecutive session to the lowest since June 2026. The mechanism is uncomfortable: the barrel is up, the terminal rate is up, the 30-year is at a 19-year high, and the dollar is falling — every rate input argued the other way. IMM data shows the positioning is rotating rather than closing: yen net shorts cut to −42,085 from −45,473, while euro net shorts deepened to −60,010 from −58,091. Forward catalyst: Wednesday's minutes; Jackson Hole, 27–29 August. | | 7. The July FOMC minutes land Wednesday into a 9–3 split. Three dissents in favour of a hike — Hammack, Kashkari and Logan — make this the most consequential scheduled item of the week, and the market has spent a fortnight pricing September out. The September hike probability read about 31% at midday Monday (Motley Fool) against roughly 50/50 a week earlier, and CME settled Friday at 32.5%. A minutes text that reads more hawkish than the vote count implies is the single cheapest way to reprice a strip that has 0.0% of any 2026 cut in it (§8). Forward catalyst: the minutes themselves, then the August CPI report on 11 September, five days before the meeting and inside the blackout. | | 8. The consumer block opens Tuesday into a tape that has punished two consecutive beats. Home Depot reports Tuesday before the open, then Lowe's, Target and TJX on Wednesday and Walmart on Thursday (§5). The setup is a July retail sales print that fell — −0.6% m/m on the Census/CNBC basis against a +0.1% consensus — and Michigan sentiment at 51.0, against Bloomberg-compiled consumer-discretionary 2026 earnings growth expectations of 31%, up from 12% in May. Forward catalyst: Home Depot, 8/18 BMO. | | --- | | | 3 · Sector Performance — August 17, 2026 | | The standing sector source could not be rendered this session, and no substitute table is published. This report's sector table is built from Finviz Groups in the Performance table view (?g=sector&v=140&o=name), which is JavaScript-gated and requires the Chrome browser tool; that tool was unreachable all session. A non-JS fetch of the same page returns stale data, and the alternative vendor boards (Investing.com sectors, Yahoo sectors) are gated the same way — the Yahoo pages that did resolve served a cached 14 August close stamped 5:15:59 p.m. EDT, which is the precise trap this report's Data Notes has warned about for six editions. No 11-group 1D/1W/YTD table is therefore published for 17 August, and the YTD compounding reconciliation is deferred to the next edition rather than run against carried-forward numbers. | | What is sourced about Monday's cross-section is worth more than a mis-dated table, and it says the breadth was worse than the index: | | • The Invesco S&P 500 Equal Weight ETF fell 0.58% against the cap-weighted index's 0.52%, and decliners beat advancers roughly 1.5-to-1 on the NYSE (Motley Fool, midday). A handful of chipmakers covered widespread losses. | | • Energy was the obvious leader by construction, with Brent +2.7% and WTI +2.6% on the expiry of the U.S.–Iran memorandum. No sector-level percentage is published because none was independently verified. | | • Semiconductors split by position in the value chain, not by end market. Up: SK Hynix +6.5% (roughly $83bn of market value added), Micron +5.9% through $1,000, Nvidia +0.9%. Down: the buyers of their output — Microsoft −2.4%, the index's largest single drag. | | • Communication services was pressured by litigation and by a 13F. Meta −3.3% into Tuesday's Oakland trial; Alphabet −0.5% despite Berkshire adding. | | • Industrials carried an idiosyncratic hit: L3Harris −4.6% on the CEO departure. | | • Consumer staples carried the day's largest S&P 500 decline: Constellation Brands −6.2% on Berkshire's full exit. | | • Nike closed the trending board at 38.93, −1.81 / −4.43% on a ~2:00 p.m. ET mark, ahead of the retail block. | | The composition read, stated without a table. This was a bond-market session expressed in equities, and the give-away is that the equal-weight index underperformed the cap-weighted one on a day when the cap-weighted index's biggest constituent was its biggest drag. That combination is unusual: normally a megacap-led decline flatters equal weight. It happened because the two forces were orthogonal — a 19-year-high 30-year yield hit everything with duration in its cash flows, which is most of the equal-weight index, while a single research note hit one megacap. Add Friday's evidence and the fortnight's rotation is legible without any sector percentages at all: the market is paying for the physical AI supply chain and selling the entities financing it. | | --- | | | 4 · Movers & Single-Name Catalysts | Up | | SK Hynix +6.5%, adding roughly $83bn in market value — the largest contributor to the memory complex's continued run, and the clearest beneficiary of a Morgan Stanley hyperscaler-capex note that the buyers of memory read as a warning (Motley Fool; non-S&P 500). Micron (MU) +5.9%, back above $1,000 per share, extending Friday's +2.42%. Nvidia (NVDA) +0.9% on disclosing it will finance up to $105bn in credit and compute for an OpenAI data centre in Pike City, Ohio — a 20-year project whose first 4.25 gigawatts arrive in 2028. | | On the ~2:00 p.m. ET trending and gainers boards (marks, not verified closes, and labelled as such): Heartflow (HTFL) +35.70% to $42.08; Applied Optoelectronics (AAOI) +15.53% to $150.28; Reddit (RDDT) +12.63% to $178.09 — the second consecutive double-digit session into its 18 August S&P 500 inclusion, following roughly +12% on Friday; Wingstop (WING) +11.05% to $126.12; MaxLinear (MXL) +10.68% to $84.84; Nu Holdings (NU) +9.33% to $15.23; HIVE Digital (HIVE) +15.23% to $3.0985. Of these only Reddit is an index event, and it is an index event twice over — it joins on Tuesday morning, the same session Home Depot reports. | | The energy complex was the day's structural winner by construction rather than by catalyst: with Brent +2.7% through $90 and WTI +2.6%, the producers, services and refiners were all bid. Individual energy single-name closes were not independently verified this session and are therefore not listed — the Investing.com Dow-components and sector boards that normally supply them are Chrome-gated. This is a gap and is recorded as one. | Down | | Constellation Brands (STZ) −6.2% — the largest S&P 500 decline of the session, after Berkshire Hathaway disclosed it had sold its entire holding in the seller of Modelo beer and Robert Mondavi wine (AP). L3Harris Technologies (LHX) −4.6% after the defence company said Christopher Kubasik stepped down as CEO and chairman following "certain conduct by Kubasik that was not consistent with the values of the Company"; the company gave few details but stated the conduct was not related to its financial reporting, controls, customer relationships or operational performance — which is the sentence that keeps this a governance event rather than an earnings event. Nike (NKE) −4.43% to $38.93 on a ~2:00 p.m. ET mark, three days into the retail block's run-up. | | Meta Platforms (META) −3.3% ahead of Tuesday's opening arguments in a federal trial in Oakland, where 29 state attorneys general allege the company designed Facebook and Instagram to get children addicted; the ~2:00 p.m. ET trending board marked it $565.89, −$23.96 / −4.06%, so the intraday low was worse than the midday print. Microsoft (MSFT) −2.4% — the S&P 500's single biggest drag — after a Morgan Stanley note argued hyperscalers are committing large upfront cash to AI data-centre builds whose payoff may take years, creating financial strain in the interim. Alphabet (GOOGL) −0.5%, notable for how little it fell: Berkshire disclosed it had increased its position alongside several homebuilders, and the stock still closed lower. | | On the ~2:00 p.m. ET losers board (marks, not closes): Bullish (BLSH) −11.24% to $24.39; Dillard's (DDS) −8.62% to $560.88, a department-store read the day before the big-box block opens; Ultra Clean Holdings (UCTT) −8.59% to $84.93, a semiconductor-equipment name falling on a day memory rallied — the same buyer/supplier split visible in §3; Duolingo (DUOL) −7.82% to $132.83; X-Energy (XE) −7.72% to $20.98. Intel (INTC) −1.97% to $102.50 on the most-active board. | | Analyst actions. No brokerage rating change with a published firm, rating and price target was independently verified for the 17 August session. The single sell-side item that demonstrably moved a large-cap was the Morgan Stanley hyperscaler-capex note, which carried no published target in the reviewed material and is reported as a thesis rather than as a rating action. Friday's actions — HSBC's Abhishek Shukla cutting Cisco to Hold with a $120 target, JPMorgan's Harlan Sur upgrading Sandisk to Overweight at $2,250, Loop Capital's Rob Sanderson initiating Reddit at $250 — remain the live sell-side context into this week and are restated rather than re-dated. | | --- | | | 5 · S&P 500 Earnings Calendar — Current & Next Week (S&P 500 components only) | | Times are ET. Earnings Whispers day pages are JavaScript-rendered and require the Chrome browser tool, which was unavailable this session, so no independent re-pull was possible. The rosters below are the 14 August verified pull, which was itself a full independent re-capture of every day page from 8/17 through 8/21 screened against the Investing.com S&P-500-component board. They are carried forward with that provenance stated, not re-presented as a fresh capture. Confirm every time against company IR before trading a date. | | | Mon 8/17 — completed. Neither the before-open nor the after-close page listed an S&P 500 reporter on the 14 August capture (ten names screened across both buckets). Nothing in Monday's tape contradicted that. | | 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 listed an S&P 500 reporter (four names screened before the open; the after-close page carried a single micro-cap). | | | The 8/24–8/28 rosters were not captured. Earnings Whispers day pages for that week require the Chrome tool and no verified alternative constituent-screened source was available. Rather than publish an unscreened list, this report states the gap: no S&P 500 roster is published for 24–28 August in this edition, and the next edition will capture the full week. What is diarised from primary sources and is not an earnings item: Broadcom (AVGO) reports on 2 September, and the Jackson Hole symposium runs 27–29 August. | | Changes vs. the prior calendar (8/14 report): | | • No additions and no removals can be asserted, because no independent re-pull occurred. Under this report's standing rule a name is treated as removed only after two consecutive absences from a capture; a session with zero captures generates neither absences nor additions. The twelve names above stand unchanged from the 14 August verified pull. | | • Toll Brothers (TOL) 4:30 and Mercury Systems (MRCY) 4:00 on 8/18 remain in the same position as Friday: absent from one capture, conservatively excluded, and not yet at two consecutive absences — this session produced no capture to count against them. They will be resolved in the next edition. | | • Index membership change, not an earnings item: Reddit (RDDT) joins the S&P 500 on 18 August, replacing AvalonBay Communities (AVB), which Equity Residential is acquiring; the combined company remains in the index as Vivmark Residential. Reddit rose about 12.6% on Monday on a ~2:00 p.m. ET mark, its second consecutive double-digit session. The rebalance flow lands on the same morning as the Home Depot print. | | • Membership caveat, restated: the constituent board used as this report's screening proxy does not carry Coherent (COHR) — retained for continuity across eight editions — nor Workday (WDAY) or Sandisk (SNDK). Confirm with company IR. | | • Timing bucket unpublished: none. Every name above carries a specific clock time from the 14 August capture. | | • What the forward calendar hands the desk. The week's entire S&P 500 reporting risk sits in a 48-hour window from Tuesday's open to Thursday's close, and six of the twelve names are U.S. consumer retailers — Home Depot, Lowe's, Target, TJX, Walmart and Ross Stores. The reaction function going in is hostile: Applied Materials was sold twice on a beat, −2.48% then −5.12%, and Tapestry fell 16.49% on a double beat and a dividend raise before stabilising. Analog Devices (8/19 BMO) is the analog read into a tape that has just paid for memory and sold the memory buyers; Keysight (8/18 AMC) is the test-and-measurement one; Deere (8/20 BMO) is the only industrial in the block. There is no S&P 500 reporter on either Monday or Friday. | | --- | | | 6 · U.S. Treasury Yields — Official Par Curve | | Basis, stated before the table. Treasury.gov's Text View is paginated and its month-specific URL requires either the Chrome browser tool or a direct fetch that the session's fetch-provenance rule blocked; the full-year CSV and XML endpoints both resolved but served copies whose last row is 14 August, so the official 17 August par curve row could not be retrieved this session. Three retrieval routes were attempted and named: the full-year Text View HTML (returned rows through early July on page one), the full-year CSV endpoint (last row 08/14/2026), and the XML feed (truncated at the 14 August entry). The table below therefore carries the official 14 August par curve as the anchor — every figure in it is an official Treasury close — and the 17 August moves are shown separately from real-time wire marks, clearly labelled. No official 17 August number is invented. | | Tenor | 14 Aug (official par) | 13 Aug | 1-day | 7 Aug | 1-week |
|---|
| 1 Mo | 3.79% | 3.79% | 0 bp | 3.79% | 0 bp | | 1.5 Mo | 3.80% | 3.79% | +1 bp | 3.79% | +1 bp | | 2 Mo | 3.81% | 3.81% | 0 bp | 3.83% | −2 bp | | 3 Mo | 3.86% | 3.87% | −1 bp | 3.87% | −1 bp | | 4 Mo | 3.88% | 3.88% | 0 bp | 3.89% | −1 bp | | 6 Mo | 3.95% | 3.94% | +1 bp | 3.96% | −1 bp | | 1 Yr | 3.98% | 3.97% | +1 bp | 4.01% | −3 bp | | 2 Yr | 4.17% | 4.15% | +2 bp | 4.19% | −2 bp | | 3 Yr | 4.24% | 4.20% | +4 bp | 4.25% | −1 bp | | 5 Yr | 4.36% | 4.32% | +4 bp | 4.35% | +1 bp | | 7 Yr | 4.51% | 4.47% | +4 bp | 4.49% | +2 bp | | 10 Yr | 4.68% | 4.63% | +5 bp | 4.65% | +3 bp | | 20 Yr | 5.25% | 5.20% | +5 bp | 5.20% | +5 bp | | 30 Yr | 5.25% | 5.21% | +4 bp | 5.19% | +6 bp |
| | The 17 August session, from real-time wire marks. | | Tenor | 17 Aug (real-time) | vs 14 Aug official par | Source |
|---|
| 10 Yr | 4.72% | +4 bp | AP, late Monday | | 30 Yr | 5.311% | +6 bp | CNBC — the highest in 19 years, and the highest since June 2007 | | 10 Yr (Asian hours) | 4.69% | +1 bp | An intra-session waypoint, shown to date the move |
| | Name the shape: a long-end-led bear steepener driven by an oil shock, and the diagnostic is inflation compensation rather than policy path. The evidence is in which end moved and why. The 30-year rose to a 19-year high on a day whose only real catalyst was Brent through $90; a higher energy price raises expected inflation over the life of a 30-year bond far more than it changes what the Fed does in September. The market's own September pricing barely budged — the hike probability read about 31% at midday against 32.5% at Friday's CME settle, a move of roughly a point and a half. A curve that steepens 2 bp at the long end while the front end prices no change is not repricing the path; it is repricing the compensation for holding duration through a supply shock. That is the same diagnostic this report named on Friday, arrived at by a different route: Friday's version was auction supply and a weak consumer, Monday's is the barrel. | | The cumulative number is the one to carry. The 10-year has gone from 3.97% before the war with Iran to 4.72% — 75 basis points, essentially all of it attributable to the energy-inflation channel rather than to Fed expectations, since the Fed has not moved and the market prices a two-thirds chance it does not move next month either. The AP's framing is the mechanism in one sentence: higher oil is "worsening inflation and upping the probability that the Federal Reserve will have to hike interest rates." The consumer transmission is already visible — the average long-term U.S. mortgage rate has jumped near its highest level in a year on the back of the 10-year. | | The bill is the separate signal and it belongs with the funding data. On the last official curve the 3-month was the only tenor that fell, easing a basis point to 3.86% while everything past two years sold off, on the day ON RRP take-up printed a third consecutive record low at $0.250bn (§9b). A bill curve easing while the coupon curve sells off is a financing signal, not a rates signal — bills are what the shrinking money-fund cushion is buying. No 17 August bill mark was retrievable, so whether that divergence persisted through Monday's oil shock is the first thing to check in the next edition. | | --- | | | 7 · U.S. Macroeconomic Calendar | Current week (Aug 17–21) | | Date | Time ET | Release | Actual | Consensus / prior | Sensitivity | Take |
|---|
| Mon 8/17 | 08:30 | Empire State Manufacturing Survey | Not obtained | — | Medium | Scheduled on the NY Fed's official August calendar; one vendor's day-ahead calendar showed the U.S. docket empty, and no actual was published in the reviewed material. The conflict is flagged, not resolved (see Data Notes) | | Tue 8/18 | 08:30 | Business Leaders Survey (NY Fed) | — | — | Low | Regional services | | Tue 8/18 | 08:30 | Imports and Exports (prices) | — | — | Low | The tariff pass-through channel; more interesting than usual with crude up 2.6% | | Tue 8/18 | 08:30 | New Residential Construction (housing starts and permits) | — | — | Medium | The single most rate-sensitive release of the week, landing into a 30-year at a 19-year high and a mortgage rate near a one-year high | | Tue 8/18 | 09:15 | Industrial Production and Capacity Utilization | — | — | Medium | The supply-side counterpart to Friday's demand miss | | Tue 8/18 | 10:00 | NAR Pending Home Sales Index | — | — | Low | Leading indicator for existing-home turnover | | Wed 8/19 | 14:00 | FOMC Minutes — July meeting | — | — | High | The week's most consequential scheduled item. Insight into the 9–3 vote split, with Hammack, Kashkari and Logan dissenting in favour of a hike | | Wed 8/19 | 10:00 | Outlook-At-Risk (NY Fed) | — | — | Low | Distributional growth and inflation risk | | Thu 8/20 | 08:30 | Initial Claims | — | — | High | Whether Thursday's 209,000 — an eight-week high — was a holiday-week artefact or a trend | | Thu 8/20 | 08:30 | Philadelphia Fed Manufacturing Survey | — | — | Medium | Pairs with Empire for the August regional composite | | Thu 8/20 | 10:00 | Reserve Demand Elasticity (NY Fed) | — | — | Low | Not a macro event, but the Fed's own read on how close reserves are to scarce, with ON RRP at $0.250bn (§9b) | | Thu 8/20 | 11:30 | Weekly Economic Index | — | — | Low | — | | Fri 8/21 | 09:45 | S&P Global Manufacturing and Services PMI (flash, August) | — | — | Medium | Private-survey first look at August; not on the NY Fed calendar | | Fri 8/21 | 12:45 | NY Fed Staff Nowcast | — | — | Low | — |
| | There is no release rated "Very high" anywhere in this week's calendar. That is itself the setup, and it is why a single geopolitical headline moved the 30-year to a 19-year high on Monday. | Next week (Aug 24–28) | | Date | Time ET | Release | Sensitivity | Note |
|---|
| Wed 8/26 | 08:30 | PCE Deflator (July) | Very high | The month's defining print. Thursday's producer report has largely written it — core services PPI accelerated from +0.1% to +0.4%, led by portfolio management +6.5%, and those lines feed the core deflator close to mechanically | | Thu 8/27 – Sat 8/29 | — | Jackson Hole Economic Symposium | High | Chair Kevin Warsh's first symposium. Macquarie's Thierry Wizman cautions he may give little insight; Warsh has been adamant about giving markets less guidance and has described his remarks as a "blank page" | | Thu 8/27 | 08:30 | Initial Claims | High | — |
| | The remaining 24–28 August items were not verified from the NY Fed's official monthly calendar this session — the calendar page requires the Chrome tool — so the block above lists only releases confirmed from primary and wire sources. The next edition will publish the full week. | | Look-ahead — the hooks, in the order they can move the Fed card. The market has spent a fortnight taking September apart and Monday barely touched it: the hike probability read about 31% at midday against 32.5% at Friday's CME settle and roughly 50/50 a week earlier. With no first-tier data until the end of the month, the sequence that can move it is: (1) housing starts and industrial production, Tuesday 8:30 and 9:15 (Medium) — the supply-side check on Friday's demand miss, and the first test of whether a 19-year-high 30-year has begun to bite the real economy; (2) the July FOMC minutes, Wednesday 14:00 (High) — three dissents in favour of a hike is already known, so the information is in the language, specifically whether the majority described the hold as a pause or as a decision, and this is the cheapest available way to reprice a strip carrying 0.0% probability of any 2026 cut; (3) initial claims, Thursday 8:30 (High) — the cleanest test of the 209,000 print, with roughly 240,000 the level that would start putting a 2026 cut into that strip; (4) the July PCE deflator, 26 August (Very high) — with a +0.4% core reading putting September back to a coin flip; and (5) Jackson Hole, 27–29 August, into a committee that had three hawkish dissents and a chair who has promised less guidance, not more. The asymmetry has flipped since Friday, and oil flipped it. A week ago the risk was that soft demand data pulled the hike out of September entirely; on Monday the barrel went through $90, the 30-year went to a 19-year high, and the inflation leg is now the live one — with August CPI on 11 September, five days before the meeting and inside the blackout, as the last word. |
| | --- | | | 8 · Fed Funds Futures & Rate Path | | Current target range: 3.50–3.75% (IORB 3.65%; EFFR 3.63%, NY Fed). The market prices hikes, not cuts: the probability of a target range below 3.50–3.75% is 0.0% at every 2026 meeting on both vendors. | | Provenance of every column in this section — read this before any number. CME FedWatch and the Investing.com Fed Rate Monitor are both Chrome-gated and the browser tool was unreachable all session. Neither the live CME meeting card nor the Investing.com per-meeting matrix could be pulled. Consequently: | | • The 17 August session read is published as a range from two independently sourced narrative marks, not as a vendor matrix. The September hike probability read about 31% at 12:26 p.m. ET (Motley Fool) and about 35% in Asian hours, with prediction markets showing roughly a 70% chance of a hold (Leveraged). Both are day-of reads; neither is a settlement snapshot. | | • The three distribution tables below are the 14 August verified pulls, carried forward with their date stamped in every heading. They are the last complete vendor matrices this report holds and they are labelled as such throughout. They are not represented as 17 August pricing. | | • No cell anywhere in this section has been estimated, chart-read or interpolated. | The headline — CME FedWatch, 16 September meeting (14 August settle, with the 17 August session read overlaid) | | Target rate (bps) | 17 AUG (session read) | 14 AUG settle | 13 Aug | 7 Aug | 14 Jul |
|---|
| 350–375 (hold) | ~69% | 67.5% | 66.1% | 55.6% | 41.9% | | 375–400 (+25) | ~31% | 32.5% | 33.9% | 44.4% | 50.0% | | 400–425 (+50) | Not published | 0.0% | 0.0% | 0.0% | 8.0% | | EASE (cut) | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
| | The 14 August column and everything to its right are published CME figures from a Current view stamped "Data as of 14 Aug 2026 05:02:44 PM CT"; contract ZQU6, expiring 30 September, mid price 96.3325, prior volume 58,460, prior open interest 233,385. The 17 AUG column is a narrative read, not a vendor table, and carries a tilde for that reason. | | The one-day and multi-day momentum read. On the session, September's hike probability moved from 32.5% to roughly 31% — about a point and a half, on a day the 30-year yield hit a 19-year high and Brent went through $90. That is the finding, and it is the mirror image of Friday's. Friday, a consumer print missing by seven tenths moved the 2026 meetings half a point and moved the 2027 strip by up to four points. Monday, an oil shock that lifted the long end 6 bp from the last official par close moved September a point and a half and — on the evidence available — left the front end otherwise intact. In both directions the market is refusing to move the near-dated meetings and doing all its work further out. The multi-day arc from CME's own columns remains the cleanest version: the September hold has gone 41.9% a month ago → 55.6% a week before Friday → 66.1% on 13 August → 67.5% at Friday's settle → about 69% on Monday, a monotonic 27-point march that has now run for four weeks without a single reversal. The +50 bp September tail, 8.0% on 14 July, is 0.0%. | The cross-check — Investing.com Fed Rate Monitor (14 August basis) | | Investing.com's card for 16 September, updated 5:45 p.m. EDT on 14 August, showed hold 67.7% (previous day 69.0%, previous week 56.6%) and +25 bp 32.3% (31.0%, 43.4%), on a September fed funds contract price of 96.330. | | The gap, reconciled and quantified. It is 0.2 percentage points — the smallest this report has recorded — and it does not reconcile the obvious way. CME's mid is 96.3325 against Investing.com's 96.330, a difference of 0.25 basis points, with CME quoting the higher price. Work the sensitivity: the 16 September meeting falls on day 16 of a 30-day contract month, so a certain 25 bp hike lifts the September average effective rate by 25 × (14/30) = 11.7 bp, which makes one basis point of ZQU6 worth about 8.6 points of headline probability. A 0.25 bp higher price should therefore put CME's hold about 2.1 points above Investing.com's; it is instead 0.2 points below. The residual is the assumed base rate. CME anchors on the prevailing EFFR of 3.63%, half a basis point above the 3.625% range midpoint, and half a basis point of base is worth roughly 4.3 points of headline probability in the opposite direction. Net the two and the residual is a fraction of a point — which is what the boards showed. The vendors are not disagreeing about the Fed; they are quoting the same contract to different conventions inside a quarter of a basis point. No 17 August version of this reconciliation is published because neither board could be read. | (a) Current-year meeting distributions — 14 August basis, current / [13 Aug] / [7 Aug] | | Meeting | 3.50–3.75 (hold) | 3.75–4.00 (+25) | 4.00–4.25 (+50) | 4.25–4.50 (+75) | Cumulative ≥ +25 | Cumulative cut |
|---|
| Sep 16, 2026 | 67.7% / [69.0] / [56.6] | 32.3% / [31.0] / [43.4] | 0.0% | 0.0% | 32.3% / [31.0] / [43.4] | 0.0% | | Oct 28, 2026 | 53.7% / [53.1] / [41.0] | 39.6% / [39.7] / [47.1] | 6.7% / [7.1] / [12.0] | 0.0% | 46.3% / [46.8] / [59.1] | 0.0% | | Dec 09, 2026 | 32.8% / [33.3] / [23.2] | 45.1% / [44.7] / [44.4] | 19.5% / [19.3] / [27.2] | 2.6% / [2.7] / [5.2] | 67.2% / [66.7] / [76.8] | 0.0% |
| | Modal ranges in bold. Contract prices: ZQU6 96.330, ZQV6 96.280, ZQZ6 96.165. | (b) Next-year meeting path — 14 August basis | | Meeting | Modal range | Prob. | Cumulative above 3.50–3.75 | Cumulative below | Contract price | Implied rate |
|---|
| Jan 27, 2027 | 3.75–4.00 | 42.8% | 73.5% | 0.0% | 96.130 | 3.870% | | Mar 17, 2027 | 3.75–4.00 | 39.0% | 79.5% | 0.0% | 96.065 | 3.935% | | Apr 28, 2027 | 3.75–4.00 | 37.4% | 81.6% | 0.0% | 96.030 | 3.970% | | Jun 09, 2027 | 3.75–4.00 | 35.8% | 82.8% | 0.0% | 95.995 | 4.005% | | Jul 28, 2027 | 3.75–4.00 | 35.8% | 82.8% | 0.0% | 95.990 | 4.010% | | Sep 15, 2027 | 3.75–4.00 | 35.7% | 81.7% | 0.6% | 95.995 | 4.005% | | Oct 27, 2027 | 3.75–4.00 | 35.4% | 80.0% | 1.4% | 96.000 | 4.000% | | Dec 08, 2027 | 3.75–4.00 | 34.9% | 77.2% | 2.8% | 96.025 | 3.975% |
| | The terminal rate is drawn by the 28 July 2027 contract at 95.990, implying 4.010%. The modal range is 3.75–4.00% at every 2027 meeting and the modal probability decays monotonically from 42.8% in January to 34.9% in December. The first non-zero probability of a target range below 3.50–3.75% anywhere on the strip is 0.6% at the 15 September 2027 meeting — thirteen months of futures with no easing priced at all. Whether Monday's oil shock pushed the terminal higher is the open question this edition cannot answer, and it is the first thing to re-read when the boards are available: Friday's template says an inflation impulse lifts the 2027 strip while leaving 2026 alone, and Monday delivered a large one. | (c) Year-end probability ladders — 14 August basis | | Year-end 2026 (9 December 2026 meeting), relative to the current 3.50–3.75%: | | Outcome | Range | Probability | [13 Aug] | [7 Aug] |
|---|
| −75 bp | 2.75–3.00 | 0.0% | 0.0% | 0.0% | | −50 bp | 3.00–3.25 | 0.0% | 0.0% | 0.0% | | −25 bp | 3.25–3.50 | 0.0% | 0.0% | 0.0% | | Hold | 3.50–3.75 | 32.8% | 33.3% | 23.2% | | +25 bp | 3.75–4.00 | 45.1% | 44.7% | 44.4% | | +50 bp | 4.00–4.25 | 19.5% | 19.3% | 27.2% | | +75 bp | 4.25–4.50 | 2.6% | 2.7% | 5.2% | | +100 bp | 4.50–4.75 | 0.0% | 0.0% | 0.0% |
| | Year-end 2027 (8 December 2027 meeting), relative to the current 3.50–3.75%: | | Outcome | Range | Probability | [13 Aug] | [7 Aug] |
|---|
| −100 bp | 2.50–2.75 | 0.0% | — | 0.0% | | −75 bp | 2.75–3.00 | 0.0% | 0.0% | 0.0% | | −50 bp | 3.00–3.25 | 0.1% | 0.3% | 0.4% | | −25 bp | 3.25–3.50 | 2.7% | 4.1% | 3.9% | | Hold | 3.50–3.75 | 19.9% | 22.4% | 17.8% | | +25 bp | 3.75–4.00 | 34.9% | 35.1% | 32.0% | | +50 bp | 4.00–4.25 | 27.5% | 25.6% | 28.3% | | +75 bp | 4.25–4.50 | 11.6% | 10.0% | 13.4% | | +100 bp | 4.50–4.75 | 2.8% | 2.2% | 3.5% | | +125 bp | 4.75–5.00 | 0.4% | 0.3% | 0.5% | | +150 bp | 5.00–5.25 | 0.0% | 0.0% | 0.0% |
| | Rounding, stated transparently. Investing.com's 14 August cards do not all sum to exactly 100.0%: the September, October and December 2026 cards and the January, September and October 2027 cards each sum to 100.0; March, June, July and December 2027 each sum to 99.9; and April 2027 sums to 100.2. CME's September headline summed to 100.0 in every one of its four columns. Cumulative figures 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. No cell in this section is estimated or chart-read, and every table is dated 14 August in its own heading. | | --- | | | 9 · Credit & Funding | (a) IG and HY credit spreads | | As-of dates, stated per row rather than implied — and they differ this session, which is itself worth recording. FRED's plain-data table endpoint served a 17 August 9:22 a.m. CDT vintage for the IG series, whose last observation is 14 August. The HY and CCC series pages served an older 14 August 8:58 a.m. CDT vintage, whose last observation is 13 August. Rather than blend them, each row below carries its own as-of date. | | Series | Level | As-of | 1-day | 1-week | YTD (31 Dec 2025) |
|---|
| IG — ICE BofA US Corporate OAS (BAMLC0A0CM) | 80 bp | 14 Aug | +1 bp (79) | +2 bp (78 on 7 Aug) | +1 bp (79) | | HY — ICE BofA US High Yield OAS (BAMLH0A0HYM2) | 271 bp | 13 Aug | 0 bp (271) | +1 bp (270 on 7 Aug) | −10 bp (281) | | CCC & lower OAS (BAMLH0A3HYC) | 1,024 bp | 13 Aug | +4 bp | +7 bp (1,017) | +139 bp (885) | | CCC minus HY differential | 753 bp | 13 Aug | +4 bp | +6 bp (747) | +149 bp (604) | | CDX IG 5y | see retrieval note | — | — | — | — | | CDX HY 5y | see retrieval note | — | — | — | — |
| | The IG series moved, and it is the first move in a fortnight. BAMLC0A0CM printed 0.80 on 14 August against 0.79 on 13 August and 0.78 on 7 and 10 August — the index had not moved more than a basis point since 3 August and had sat in a 78–79 band for nine consecutive prints. 80 bp is the widest since 24 July. It is one basis point and it would be noise on its own; it is not noise in the company it keeps, because it printed on the same session as a 36% order-attrition week and a rating action taken against a financing structure. No 17 August index credit spread exists at any vendor — FRED's one-business-day lag means the 17 August print publishes on 18 August — so Monday's oil shock is not yet in these numbers, and that is the single most important caveat in this section. | | CDX retrieval note — the six-step ladder was worked to the limit the toolset allowed, and the level remains unobtainable. (1) Bloomberg in Chrome and (2) WSJ Market Data — both require the Chrome browser tool, which was unreachable for the entire session across three connection attempts. This is a harder failure than Friday's: Friday both pages rendered and were verified programmatically to contain no CDX string; Monday they could not be opened at all. (3) Cbonds / ICE / S&P Global: Cbonds' CDX.NA.IG 5Y page was confirmed entitlement-gated on Friday, displaying the level as three masking asterisks in place of a number, and nothing changed that. (4) FT Markets Data and Reuters credit wraps: no instrument page for the index and no wire wrap quoting a level appeared in the reviewed material. (5) TradingView / Barchart / CME CDS index product pages: these return methodology and index-family documentation; the most recent third-party level found anywhere remains an undated-context 81 bp reference from 27 July, and under this report's standing rule an undated digest number is not a CDX level and is withheld. (6) Cash-market proxy, labelled as such: no verified 17 August HYG or LQD close was obtainable either, because the same vendor boards are gated. Quoting conventions restated so no reader mis-signs the field: CDX IG 5y is quoted in basis points of spread; CDX HY 5y in price points, where a rising price means tightening credit spreads. | (b) Money-market and funding plumbing | | NY Fed reference rates publish at roughly 8:00 a.m. ET for the prior business day. The NY Fed's public rates API resolved cleanly this session and returned 13 August fixings as its latest for the overnight series, with the SOFR averages and index carrying a 14 August effective date. | | Measure | Level | Detail |
|---|
| SOFR | 3.62% | 13 Aug. Volume $2,932bn; 1st percentile 3.59%, 25th 3.60%, 75th 3.67%, 99th 3.70% | | EFFR | 3.63% | 13 Aug. Volume $106bn; 1st percentile 3.60%, 25th 3.62%, 75th 3.63%, 99th 3.65% | | OBFR | 3.63% | 13 Aug. Volume $236bn; 1st percentile 3.53%, 99th 3.68% | | TGCR | 3.60% | 13 Aug. Volume $1,207bn | | BGCR | 3.60% | 13 Aug. Volume $1,232bn | | SOFR − IORB | −3 bp | IORB 3.65%; SOFR still trades through the administered rate | | 30-day average SOFR | 3.63617% | 14 Aug effective date; 90-day 3.63113%, 180-day 3.66204%; SOFR index 1.25477279 | | ON RRP take-up (14 Aug) | $0.250bn | A third consecutive record low, from $0.450bn on 13 August and $0.725bn on 12 August | | Reserve balances (week to 12 Aug) | $2,944.1bn | −$198.6bn from the 15 July peak of $3,142.7bn; no new weekly print available | | 3-month bill (par, 14 Aug) | 3.86% | −1 bp, the only tenor on the last official curve that fell (§6) |
| | Read this with §6, because they are the same story. The ON RRP facility has effectively emptied — $0.250bn is a rounding error against a $2.4tn peak — and take-up printed a fresh record low on three consecutive sessions. Reserves are $198.6bn below the July peak. And yet there is no daily stress anywhere in the fixings: SOFR at 3.62% is 3 bp through IORB, the 1st-to-99th percentile band is 11 bp, tri-party and broad general collateral both clear at 3.60% on $1.2tn a day, and EFFR sits mid-range with a 5 bp band. The adjustment is showing up in the bill curve rather than in the overnight rate, which is what happens when the marginal cash that used to sit at the facility has to buy something. There is no cushion left for September quarter-end, and the Fed publishes its own Reserve Demand Elasticity read on 20 August at 10:00 (§7) — now the most informative scheduled item on the funding question. No month-end or quarter-end distortion is present in the 13 August fixings; mid-month is the quiet part of the cycle, which is precisely why the RRP number is notable. | (c) Rates volatility and swap spreads | | No 17 August MOVE value was obtainable. The Investing.com series is delayed, published no 14 August value at the time of Friday's writing, and its page requires the Chrome tool. The last verified reading is 69.23 on a 13 August vintage, down 2.86 points / −3.97% from 72.09 on 12 August. Against a VIX of 15.25 on a ~2:00 p.m. ET 17 August mark that gives MOVE/VIX of about 4.54×, which pairs two dates four sessions apart and should not be traded on; it is quoted only for continuity with prior editions, which ran 4.86× and 4.93× on the same mismatched basis. | | What can be said without a date mismatch is the direction, and it reversed. Equity vol rose 7.01% on Monday after falling to a seven-month low on Friday. Rates vol, on its last available reading, was falling 4.0% into a week in which the 30-year auction cleared at the highest yield since 2001 and the 10-year at the highest since 2007 — and the 30-year has since gone to a 19-year high. If the MOVE has not risen on a 6 bp long-end move to a two-decade high, the rates-vol market is mispriced against its own underlying, and that is the specific thing to check when the board is readable. 2y, 10y and 30y swap spreads could not be retrieved — the Bloomberg rates board, the WSJ money-rates table and the FT markets search are all Chrome-gated and none was reachable. This is a gap and is recorded as one rather than filled, for the second consecutive edition. | (d) Issuance, leveraged loans and private credit | | The primary market's indigestion is the most interesting standing fact in credit and nothing on Monday resolved it. Per Bloomberg's Friday reporting, investors pulled about 36% of their initial orders for high-grade deals last week after final pricing was squeezed — twice the prior week's rate, and well above the 22% average for the year — with Thursday's attrition reaching 45%, a record for 2026. The context is volume: $136bn priced in the preceding fortnight, August on track for its busiest ever, $1.4tn of IG sold year to date, 9% above the pace of 2020's record year, and global issuance reaching $5tn more than a month faster than last year's record. This week's expected supply is only about $20bn, and the market usually goes quiet until after Labor Day — so the next real test of the bid is the post-Labor Day burst, not this week. | | Fund flows turned for the first time in six weeks, and they turned in tech. Global technology equity funds posted $1.7bn of outflows, snapping a six-week inflow streak, while global equity funds took in $18.62bn in the week to 12 August — a twelfth consecutive weekly inflow, led by Europe at $13.52bn and Asia at $4.13bn. Money is still arriving; it is arriving somewhere other than U.S. technology. Separately, the GDX gold-miner ETF has attracted $419m in August, on track for its strongest monthly intake since February. | | Private credit and vendor financing is where the equity market keeps pricing the risk. Monday added the largest data point yet and it is not a rating action: Nvidia disclosed it will finance up to $105bn in credit and compute for a single OpenAI data centre. That is vendor financing at a scale that changes the counterparty question — the supplier is now the lender. Set it beside Friday's Broadcom episode, where Bank of America cut the issuer and bond ratings over the AI XPV platform built with Apollo and Blackstone, and the two large caps on either side of it moved −5.95% and −3.61%. On the leveraged loan side, the most recent published Morningstar LSTA index analysis available remains June 2026, when the index returned 0.08%, its weakest month since February; no verified August level was obtainable and it is flagged rather than estimated. Named private-credit watch items: the XPV vehicle; Apollo and Blackstone as its sponsors; and now Nvidia's own credit commitment, which is a vendor-financing exposure sitting inside an equity that trades on a demand multiple. | | The take — credit has not yet seen Monday, and that is the most important sentence in this section. The FRED series stop at 14 August, so the 19-year-high 30-year, the $90 Brent and the 7% VIX move are all absent from every credit spread quoted above. What the numbers do show is a configuration that had already begun to move at the margin: IG at 80 bp, its first widening in a fortnight and the widest since 24 July; HY at 271 bp, still 10 bp inside January; the CCC credit spread at 1,024 bp against 885 bp on New Year's Eve, with the CCC-minus-HY differential at 753 bp against 604 bp. Index credit is priced for no defaults and the tail is priced for 149 bp more of them than it was in January. Against that, equity vol has finally moved — VIX +7.01% — so for the first time in this reporting window the credit-versus-equity-vol divergence narrowed from the equity side rather than the credit side. That is the wrong way round for anyone who owns the divergence as a trade: you wanted credit to come to vol, not vol to go to credit. What would break it, specifically: an IG print on 18 August that extends 80 bp rather than retracing it, which would make Friday's widening a trend rather than a tick; a 36% order-attrition rate persisting into the post-Labor Day supply burst; a second issuer put on negative watch for a financing structure rather than for earnings; or September quarter-end arriving with ON RRP at $0.250bn and no cushion (§9b). Note what would not break it: another weak consumer print — Friday delivered one and IG did not move a basis point that day. |
| | --- | | | 10 · FX — Levels and Moves | | Quote basis — read this before the table, because it is unusually weak this session. 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 and GBP/USD it means the dollar weakened. The TradingEconomics board and the Investing.com currency pages are both Chrome-gated and neither could be rendered. The levels below are therefore Asian-hours and early-European marks captured at roughly 07:08 UTC (3:08 a.m. ET) from a dealer-desk review, cross-referenced against a TradingEconomics narrative summary published later in the day. They are not New York closes and are not presented as such. The Asian crosses this report normally carries — USD/KRW, USD/TWD, USD/CNY, USD/INR, USD/MXN, USD/RUB — had no verified 17 August mark from any reachable source and are omitted rather than carried forward at Friday's levels. | | Pair | Level (basis stated) | %Chg | Note |
|---|
| DXY | ~99.40–99.50 | −0.14% | A third consecutive decline, and TradingEconomics marks it the lowest since June 2026. Asian-hours basis | | EUR/USD | 1.1585–1.1588 | +0.14% | The strongest since mid-June. BBH marked the euro "firm at $1.1578" later in the session | | USD/JPY | 159.07–159.09 | −0.16% | A $2.3bn option expiry at 159.00 sat on the price. BBH later marked 159.15 | | GBP/USD | 1.3553 | +0.1% | Holding firm on the broad dollar theme; net spec shorts reduced to −56,221 | | AUD/USD | Not published | — | No verified 17 August mark; net spec shorts increased to −39,223 from −33,190 | | Bitcoin (reference) | $62,956–63,000 | −0.19% | Range-bound, with $63,400–63,500 as short-term resistance |
| | The take — the dollar fell on a day when every rate input said it should rise, for the fourth session running, and the positioning data now explains why. Run the inputs: Brent +2.7%, the 30-year to a 19-year high, the 10-year +4 bp to 4.72%, and a September hike probability that barely moved. On a rate-differential model that is an unambiguous dollar bid. The dollar index fell for a third consecutive session to its lowest since June. The explanation is not macro, it is inventory: speculative long USD positioning stands at +$48bn, more than quadruple its level three months ago and the most crowded the trade has been since 2015 — beyond two standard deviations from its mean. When a position is that stretched, the marginal flow is liquidation, and liquidation is price-insensitive to the fundamentals that created it. The unwind is rotating rather than closing, which is the detail that makes it tradeable: IMM data shows yen net shorts cut to −42,085 from −45,473 while euro net shorts deepened to −60,010 from −58,091 and Aussie shorts jumped to −39,223 from −33,190. | | The contrarian cross this session is EUR/USD, and it is contrarian in an unusual way. The euro made its strongest level since mid-June while euro net speculative shorts got bigger. Those two facts are only compatible if the rally is being driven by spot dollar selling rather than by euro buying — real-money and corporate liquidation of dollar longs going through the most liquid pair on the board, while the leveraged community adds to euro shorts against it. That is a crowded-versus-crowded standoff, and it usually resolves violently. The practical read: EUR/USD strength here is not a European story and should not be traded as one. Nomura's call for EUR/JPY at 190 by end-September and JPMorgan's year-end USD/JPY target of 164 — the latter arguing explicitly that even faster Bank of Japan tightening toward 2.00% by end-2027 will not lift the yen, citing a negative correlation between hike expectations and JPY strength — are both cross expressions of the same view, which is that the dollar's weakness is idiosyncratic and the yen's is structural. The yen is the one to watch on Wednesday: shorts have already been cut by 3,388 contracts, and a hawkish FOMC minutes text would put them back on. | | --- | | | 11 · Commodities | | Basis, stated before the table, and it is mixed this session — read the basis column, not just the level. The Investing.com per-contract historical boards, which normally supply the official front-month settle, are Chrome-gated and were unreachable, as was the TradingEconomics commodity board that supplies the week/month/YTD columns. Crude is sourced from wire settlement reporting (AP and CNBC) and is a true settle. Metals are dealer-desk and quote-board marks at the timestamps shown and are not settles. Week, month and YTD columns are omitted entirely rather than carried at Friday's values. | | Commodity | 17 Aug level | Chg | Basis | Driver |
|---|
| Brent (front, ICE) | $90.87 | +2.7% | Settle (AP) | Through $90. The 17 June U.S.–Iran memorandum expired unrenewed; Iran's foreign ministry ruled out an extension | | WTI (front, NYMEX) | $84.50 | +2.6% | Settle (CNBC) | From Friday's $82.40 settle the arithmetic is +$2.10 / +2.55%; CNBC's +2.6% is quoted as published | | Gold (Comex front) | $4,477.20 | +$39.90 / +0.90% | ~2:00 p.m. ET quote board | Rose with yields — the inflation-hedge signature, not the real-rate one | | Gold (Comex front, Asian hours) | $4,451.30 | +$14.00 / +0.32% | ~07:08 UTC dealer mark | Both bases shown so the intraday path is visible, not to double-count | | Gold (spot) | $4,389.83 | +0.32% | ~07:08 UTC dealer mark | Spot-versus-futures basis of roughly $61, consistent with carry at these rates | | Silver (spot) | $65.19 | +0.8% | ~07:08 UTC dealer mark | Outperformed gold on the same session | | Copper (Comex front) | Not published | — | — | No 17 August mark was obtainable from any reachable source; withheld rather than carried at Friday's $6.6065 | | Natural gas, products, platinum, palladium, EU gas, uranium | Not published | — | — | These rows come from the TradingEconomics board, which is Chrome-gated. Omitted rather than dated to Friday |
| | The take — this was a supply shock with a date on it, and the market had been discounting the date. The 17 June memorandum of understanding carried a 60-day life and everyone knew when it ended. It ended Monday with nothing signed, Iran's foreign ministry ruled out talks to extend it, and Trump said he was in no rush and separately threatened to bomb Oman — the country that has been mediating. Brent rose 2.7% to $90.87 and WTI 2.6% to $84.50. The physical evidence is more striking than the price. Shipping through the Strait of Hormuz has essentially stopped: three vessels crossed on Sunday against roughly 130 daily before the war began in February, and only six crossed on Friday. Iran's position, per its deputy foreign minister, is that any reopening happens on Iran's terms; the IRGC's political deputy said Iran's actions have so far been defensive but could become offensive. Against a 98% reduction in transit volume, a $90 Brent print is not an escalation premium — it is a market that still believes the strait reopens. | | The positioning read, and the range is the tell. Brent has spent the past month zigzagging between $72 and $102 as deal hopes rose and fell (AP). That is a 40% peak-to-trough range in a benchmark commodity inside thirty days, which tells you the marginal price is being set by headline traders rather than by physical balances — and it is why a scheduled expiry moved it 2.7% while a 98% collapse in Hormuz transit has not moved it above the July high. Two things cut against the bid from here: OPEC has cut its global oil demand growth forecast, and Trump has told Americans directly to expect higher gasoline prices, which is an administration signalling it has limited ability to influence the situation rather than one preparing an intervention. | | Gold rose with yields, and that is the whole precious-metals story. The metal gained 0.90% on the Comex front-month quote board on a session when the 10-year rose 4 bp and the 30-year hit a 19-year high — both unambiguously gold-negative on a real-rate model. Gold rose anyway, which is the inflation-hedge signature, and it is consistent with the transmission in §6: the barrel is what is moving the long end, and the barrel is what gold is hedging. Silver outperformed at +0.8% against gold's Asian-hours +0.32% on the same basis, which adds an industrial leg to a monetary move. The flow data corroborates: GDX has taken $419m in August, on track for its strongest month since February, and UBS expects gold at $5,000 by early 2027. The caveat that has to travel with the enthusiasm — the two gold marks in the table are four hours apart and neither is a settle, so the 0.90% figure is an intraday reading and the daily change on a settlement basis is not established. | | --- | | | 12 · Trading Views | | 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. | | A note on marking this edition. Several positions below cannot be marked precisely because the single-name and vendor boards that price them were unreachable. Where a mark is unavailable it is stated as unavailable rather than estimated, and the position is held at its prior size until it can be marked properly. A trade you cannot mark is a trade you cannot add to. | 1. The rates trade — hold the winner, and the oil shock is the reason the stop stays where it is | | Mark on the prior edition's rates idea. The book is long ZQZ6 against short ZQZ7, DV01-matched one-for-one at half of half size, entered at 96.165 / 96.060 — a spread of 10.5 bp — and Friday took it to 14.0 bp for a gain of +3.5 bp, or $145.85 per contract pair at $41.67 of DV01 per basis point. Monday's mark is unavailable: neither CME nor Investing.com contract prices could be read, so the 17 August spread is not published. What is known is directional and it favours the position — the front leg's driver (September pricing) moved about a point and a half toward a hold, while the back leg's driver (2027 inflation compensation) received a $90 Brent print and a 19-year-high 30-year. That is the exact configuration the spread is long. | | Modal path, base case and the tails. Modal path on the last complete matrices: hold on 16 September (CME 67.5% at settle, about 69% Monday, ease 0.0%); hold on 28 October at 53.7% modal with the cumulative hike at 46.3%; one 25 bp hike delivered by 9 December, 3.75–4.00% modal at 45.1% with the cumulative at 67.2%; terminal 4.010% drawn by the 28 July 2027 contract at 95.990; year-end 2027 modal 3.75–4.00% at 34.9%, with the first non-zero easing probability anywhere on the strip at 0.6% in September 2027. Base case: demand is slowing (retail sales −0.6%, Michigan 51.0) while the supply-side inflation impulse has just been re-lit by the barrel. That combination delays the hike and raises the terminal — the shape the position is long. Tail one, dovish: claims through roughly 240,000 on 20 August would price a 2026 cut into a strip showing 0.0% at every meeting; the front leg gains but the back leg gains more and the spread compresses. Tail two, hawkish: a hawkish minutes text on Wednesday puts September back toward a coin flip — the front leg loses and the back leg is unchanged. Both tails compress the spread, which is why the stop matters more than the target. | | Expression: long ZQZ6 against short ZQZ7, DV01-matched one-for-one at $41.67 per basis point per contract. Catalyst: FOMC minutes 8/19 14:00; initial claims 8/20 8:30; the 26 August PCE deflator; Jackson Hole 27–29 August. Invalidation, unchanged: the spread back through 11.0 bp; or a core PCE deflator at or below +0.2% m/m on 26 August; or any 2026 meeting showing a non-zero cut probability. Sizing: unchanged — do not add, and do not add specifically because the position cannot be marked today. | 2. Own the crack, not the barrel — hold at small; unmarked | | Expression: long a basket of Valero, Phillips 66 and Marathon Petroleum against a short in front-month WTI, sized so the futures leg matches the basket's crude beta. Mark: unavailable — no verified 17 August close for any of the three refiners was obtainable. Directionally the short leg lost 2.6% on the crude move, so the basket needed to beat that to break even, and on Friday's evidence the equity expression had stopped tracking the crack. What is intact: the crack thesis, which a supply disruption at the crude end mechanically supports. What is unresolved: whether the refiner equities participated, which is precisely what broke on Friday. Catalyst: the EIA weekly report; any Hormuz development. Invalidation, unchanged: a confirmed Hormuz reopening; or a distillate build above 3m barrels. Sizing: small, unchanged. | 3. Sell the rumour premium, buy the delivered miss — retire it | | Expression: long an equal-weight pair of Cisco and Applied Materials against short an equal-dollar basket of GoDaddy, CoStar, Paycom and Fair Isaac. Mark: unavailable. Decision: retire the idea at its last mark rather than carry it unmarked into a third session. The construction warning from Friday stands — the short basket collapsed 3.08% exactly as the thesis required and the pair still lost money because the long leg had its own idiosyncratic news. An idea whose thesis works and whose expression loses is a construction problem, and a construction problem should not be carried through a data blackout. Closed, honestly, at a small loss. | 4. Own downside in the consumer block, not in the index — hold into Tuesday, unchanged | | Expression: a one- to three-week put spread on a consumer-discretionary basket or the sector ETF, struck beneath Thursday's level and financed by selling the wing. Mark: the structure gained on Monday's 0.52% index decline and on Nike −4.43% and Dillard's −8.62% on intraday marks, though no sector-level percentage is available to price it precisely. Neither invalidation triggered. The point of the trade is the next 48 hours, not the last one: six consumer names report between Tuesday's open and Thursday's close. Catalyst: Home Depot 8/18 BMO; Lowe's, Target, TJX 8/19 BMO; Walmart, Ross Stores 8/20. Invalidation, unchanged: a control-group print above +0.5% m/m, or Home Depot rising more than 3% on its print. Sizing: small, defined-risk, premium-at-risk only — do not roll it past 8/20; the thesis expires with the block. | 5. Long gold against the Swiss franc — hold at a quarter, and the thesis got its cleanest test yet | | Expression: long spot gold funded in CHF, notional-matched, at quarter size. Mark: the long leg gained 0.90% on the Comex quote board (+0.32% on the Asian-hours spot basis); the franc leg is unmarked because no verified USD/CHF level was obtainable, so the pair's net is not published. Thesis, confirmed on its narrowed terms: this position was reduced to the inflation-hedge reading only, and Monday was the cleanest possible test of it — the 10-year rose 4 bp and the 30-year hit a 19-year high, both gold-negative on a real-rate model, and gold rose anyway on an oil shock. That is exactly what an inflation hedge is supposed to do and exactly what a real-rate proxy is not. Catalyst: the 26 August PCE deflator; any further Hormuz escalation, which is gold-positive and franc-positive simultaneously and would make this a poor hedge. Invalidation, unchanged: spot gold below $4,300; or the franc strengthening more than 0.75% against the dollar in a session where gold is flat or lower. Sizing: a quarter, unchanged — one confirming session on an unmarkable cross does not earn capital. | 6. Protection on the CCC cohort funded in IG — hold; the funding leg finally moved | | Expression: long CCC-exposed credit protection (or short a levered-loan/CCC-heavy vehicle) against long IG cash. Mark: the CCC leg is unchanged at its 13 August reading of 1,024 bp with the differential at 753 bp. The news is in the funding leg: IG widened to 80 bp on 14 August, its first move in nine prints and the widest since 24 July. That is a cost to this structure, not a benefit — the IG cash leg is what funds it — and it is the first time in this reporting window the funding leg has gone the wrong way. Thesis, unchanged but now on notice: if IG has begun to move, the "IG does nothing" assumption that makes the funding free needs re-testing on the 18 August print. Catalyst: the 18 August FRED update, which will carry the first post-oil-shock credit spread; the post-Labor Day supply burst against a 36% order-attrition rate. Invalidation, tightened: the differential back through 735 bp, or IG widening beyond 85 bp, which would make the funding leg the dominant P&L driver and break the structure's logic. Sizing: small, reduced from three-quarters back to a half of the small allocation — a funding leg that has started moving deserves less notional, not more. | 7. On-balance-sheet AI funding against off-balance-sheet AI funding — hold, and Monday complicated it usefully | | Expression: long AMD against short Blackstone, dollar-neutral, small. Mark: unavailable — no verified 17 August close for either name. What Monday added: Nvidia's commitment to finance up to $105bn in credit and compute for a single OpenAI data centre, which sits awkwardly across the thesis. It is on-balance-sheet in the sense that it is disclosed and attributable, and off-balance-sheet in the sense that it converts a chip vendor into a credit provider to its own customer. The equity moved 0.9%. The honest reading is that the market has not decided how to price vendor financing at all, which is a reason to keep the pair small rather than a reason to abandon it. Catalyst: Broadcom's 2 September print; the post-Labor Day IG calendar and whether attrition normalises toward the 22% year average; any rating action referencing a financing vehicle rather than earnings. Invalidation, unchanged: BofA restoring Broadcom's ratings; or AMD's IG spread widening more than 25 bp from new issue; or the pair losing more than 5% in five sessions. Sizing: small, unchanged. | 8. New — long the memory suppliers against the hyperscaler buyers | | Expression: long an equal-weight pair of Micron and SK Hynix against short Microsoft, dollar-neutral, small. Thesis: Monday priced one research note in two directions and the split was clean. Morgan Stanley's hyperscaler-capex warning — heavy upfront cash, payoff years out, strain in between — took Microsoft down 2.4% and made it the S&P 500's largest single drag, while the suppliers of the components that capex buys rose: SK Hynix +6.5%, adding roughly $83bn of market value, and Micron +5.9% back through $1,000. Both reactions are rational and they are the same view: if the spend is committed regardless of the payoff timing, you want to own the seller's revenue and not the buyer's cash flow. The corroborating macro is real trade data, not sentiment — Singapore's July non-oil domestic exports rose 24.2% year-on-year, a fourth straight month above 20% on AI-linked electronics. The second-order detail: global technology equity funds saw $1.7bn of outflows, snapping a six-week streak, while global equity funds took in $18.62bn — so the money leaving tech is leaving the buyers, and the suppliers are being funded out of the same pot. Catalyst: Nvidia's next print; any hyperscaler capex guidance revision; Micron's next earnings. Invalidation: a hyperscaler cutting its capex plan, which takes the supplier revenue with it and makes the short leg the winner for the wrong reason; or a memory pricing rollover, evidenced by a contract-price decline in the monthly DRAM/NAND fixings; or the pair losing more than 5% in five sessions. Sizing: small — one session of evidence, and note the long leg includes a non-U.S. listing with its own FX and session-timing basis. | | Volatility note. VIX 15.25, +7.01% on a ~2:00 p.m. ET mark — the first meaningful bid in equity vol in weeks, three sessions after a 14.25 close that was the lowest since 29 December. The important part is what it bid on. Friday's consumer print missed by seven tenths and the VIX fell 2.60%; Monday's oil shock moved it +7.01% on a smaller index decline. Equity vol will not pay for a demand problem and will pay for a supply problem — which is a coherent view, because a demand miss is disinflationary and multiple-supportive while a supply shock raises the discount rate. At 15.25 one-month implied prices a daily move of about 0.96%, against a realised session of 0.52%, so the short-vol carry is still positive but the cushion is thinner than it was. The specific objection is no longer calendar, it is correlation: the market has just demonstrated that a single headline can move the barrel 2.7%, the 30-year to a 19-year high and the index half a percent simultaneously. That is a regime in which index vol and rates vol should both be bid, and on the last available reading the MOVE was falling (§9c). If you want to own one of the two, own the one that has not moved. | | --- | | | 13 · Risk Map | | The crowded consensuses, with the numbers that would break them. | | 1. "The dollar long is the right side of the rate differential." It is the most crowded FX position since 2015 — speculative longs at +$48bn, more than quadrupled in three months, beyond two standard deviations — and it has now fallen for three consecutive sessions into a tape that handed it a 19-year-high 30-year, a +4 bp 10-year and a 2.7% Brent rally. Every one of those is dollar-positive on the model the position is built on. The stress test: a hawkish FOMC minutes text on Wednesday that fails to bid the dollar would confirm the unwind is inventory rather than macro, and the next stop is technical rather than fundamental. Note the rotation already visible: yen shorts cut to −42,085, euro shorts deepened to −60,010. | | 2. "Oil is priced for the disruption." Brent at $90.87 against a Strait of Hormuz carrying three vessels on Sunday versus roughly 130 daily before February — a ~98% collapse in transit — and a month-long range of $72 to $102. The price is not at the top of its own recent range despite the physical situation being at its worst. The stress test: Iran's stated position is that reopening happens on its terms and that the strait stays shut until the U.S. "accepts defeat"; the IRGC says actions could turn offensive; Trump has threatened to bomb the mediator. Against that, OPEC has cut its demand growth forecast — so the two sides of the argument are a supply collapse and a demand downgrade, and the market is currently splitting the difference at $90. | | 3. "Credit is fine." IG at 80 bp — but that is a 14 August number and it is the first widening in nine prints and the widest since 24 July. HY at 271 bp, 10 bp inside January. The CCC credit spread at 1,024 bp against 885 bp on New Year's Eve, differential 753 bp against 604 bp. The stress test, and it has a date: the 18 August FRED update carries the first credit spread struck after the oil shock and the 19-year-high 30-year. If IG extends past 80 rather than retracing, the fortnight of immobility is over. Add the 36% order-attrition rate against a 22% year average and the Broadcom rating action taken about a financing structure rather than about earnings. | | 4. "The AI trade is one trade." Monday proved it is at least two, moving in opposite directions on the same research note: SK Hynix +6.5% and Micron +5.9% against Microsoft −2.4%, the index's largest drag. And Nvidia rose 0.9% on committing up to $105bn. The stress test: vendor financing at that scale means the chip supplier now carries customer credit risk, and no rating agency has yet said anything about it. The theme's funding leg is where it breaks first — that has now been the case for four consecutive sessions. | | 5. "The consumer is fine because earnings are fine." Bloomberg-compiled expectations have S&P 500 consumer-discretionary 2026 earnings growth at 31%, up from 12% in May, and the index is tracking roughly 50% EPS growth for the spring quarter, per FactSet — the best in five years. Against that: retail sales −0.6% on the Census basis, Michigan 51.0, employers cutting more jobs than they added last month, and Trump telling Americans to expect higher gasoline prices. The stress test: six of these companies report inside 48 hours starting Tuesday, into a tape that took 16.5% out of Tapestry on a double beat and a dividend raise. | | 6. "Volatility is cheap because nothing happens." It just stopped being true: VIX +7.01% to 15.25. The stress test is now the other way round — the market has repriced equity vol for a supply shock while, on the last available reading, rates vol was falling 4.0% into a 30-year at a two-decade high (§9c). One of those two markets is wrong, and the 30-year is the one with the auction calendar. | | The two-sided geopolitical tape. Escalation: the 17 June memorandum expired unrenewed; Iran's foreign ministry ruled out extension talks and a senior official told Reuters Tehran is going "fully offensive" and would use its military to break the U.S. naval blockade if talks fail; the IRGC's political deputy said actions so far have been defensive but could become offensive; deputy foreign minister Kazem Gharibabadi said any Hormuz reopening is on Iran's terms; Trump told Fox News he is in no rush and threatened to bomb Oman, the mediator; Hormuz transit is down roughly 98%. De-escalation, or at least containment: Brent at $90.87 is well below its $102 July high, gold's move was under 1%, and the VIX at 15.25 is still historically low. The asymmetry is that escalation is priced in exactly one asset — the barrel — and the barrel is not priced for the physical situation that already exists. | | Structural watch items. (1) Term premium, not policy, is setting the long end, and oil is now setting term premium. The 30-year at a 19-year high, the 10-year 75 bp higher than before the war at 4.72% from 3.97%, and the average long-term mortgage rate near a one-year high as the direct consumer transmission. (2) The funding cushion is gone and there is still no daily stress. ON RRP $0.250bn, a third consecutive record low; reserves $198.6bn below the July peak at $2,944.1bn; SOFR 3 bp through IORB; repo clearing $1.2tn a day. September quarter-end will be the first in years with no buffer, and the Fed's Reserve Demand Elasticity read lands 20 August. (3) Vendor financing is the new off-balance-sheet risk, and it is now measured in nine figures: Nvidia up to $105bn to one customer, on top of the XPV vehicle that cost Broadcom and Blackstone 5.95% and 3.61% in a session. (4) Governance and litigation are producing large-cap moves without any change to cash flow: L3Harris −4.6% on a CEO departure for conduct, Meta −3.3% into a trial brought by 29 state attorneys general, Constellation Brands −6.2% on a 13F. (5) A new Fed chair's first Jackson Hole, 27–29 August, from a chair who has described his remarks as a "blank page" and has been adamant about giving markets less guidance — into a committee that split 9–3 with three dissents in favour of a hike. | | What VIX is and is not pricing. At 15.25 it prices a 0.96% daily move — up from 0.90% on Friday, and the increase is the first honest thing equity vol has done in a fortnight. It now prices a geopolitical supply shock. It still does not price the second-order consequence of one, which is that a barrel through $90 puts the inflation leg of the Fed debate back in play five days before the July minutes and nine days before the PCE deflator, into a strip carrying 0.0% probability of any 2026 cut and a 19-year-high 30-year. The market has learned to hedge the headline. It has not yet hedged what the headline does to the discount rate. |
| | --- | | | Sources Associated Press (Stan Choe market wrap, via TV News Check); CNBC (Treasury yields and FOMC minutes preview; oil prices, Iran war and Strait of Hormuz); Yahoo Finance (live blog and quote boards, with cached-value caveats noted throughout); TheStreet (Stock Market Today, 17 August); The Motley Fool via Yahoo Finance (midday market and single-name detail); Leveraged Financial Market Review 17.8.26 (Asian-session FX, commodities, positioning and calendar); U.S. Department of the Treasury Daily Treasury Par Yield Curve Rates (full-year CSV and XML endpoints); Federal Reserve Bank of St. Louis FRED (ICE BofA OAS series BAMLC0A0CM, BAMLH0A0HYM2, BAMLH0A3HYC); Federal Reserve Bank of New York reference-rates API (markets.newyorkfed.org/api/rates/all/latest.json); Bloomberg, WSJ, Investing.com, CME Group FedWatch, Finviz, TradingEconomics and Earnings Whispers attempted and unreachable this session — all are Chrome-gated and the browser tool was down; each is named at the point of failure in the relevant section. Prior-edition figures carried forward with their dates stamped come from US_CrossAsset_Daily_2026-08-14, this report's own verified pull. | | Full Data Notes & Conflicts and the complete categorized Source Links are in the companion file US_CrossAsset_Daily_2026-08-17_DataNotes.txt. | | U.S. Stock, Fixed Income & Cross-Asset Closing Daily — generated 17 August 2026. Prepared for institutional investors. Not personalized investment advice; verify independently before acting. |
|