Economic Risk Picture 2026

Update June 17, 2026: worse. Compared to the original/latest previous text: the risk picture is worse. The Guardian reports Trump claiming the Iran deal is “all signed”, but the same news flow still shows skepticism, Israel/Lebanon risk, and no proven physical normalization through Hormuz. That strengthens the core point: headlines exist, but the molecule test is still not passed. Guardian: Iran deal claims

Sources: Guardian: Iran deal claims IEA: Strait of Hormuz

Update June 24, 2026: same. Compared to June 17: the picture is mixed. More traffic and lower paper prices help, but EIA shows Cushing down to 20.03 million barrels for the week ending June 12. The molecule test is still not passed until flows show up in inventories, insurance, and delivery.

Sources: EIA: Cushing stocks IEA: supply readjustment after Hormuz shock

Update July 1, 2026: same. Compared to June 24: the molecule test is still not passed. The latest published EIA data still show Cushing down at 18.96 million barrels for the week ending June 19, and the WPSR page still lists that as the latest release before the July 1 update. Headlines and lower paper prices therefore do not prove physical normalization yet.

Sources: EIA: Cushing weekly stocks EIA: Weekly Petroleum Status Report IEA: Strait of Hormuz

Update July 29, 2026: worse. Compared to July 1: June's partial recovery did not hold. The IEA says Gulf oil exports recovered to 16.1 million barrels per day in June but remained well below the pre-war 24 million barrels per day; renewed hostilities then cut visible Hormuz traffic to just 11 ships on July 12, according to S&P Global. On July 29, Iran rejected Oman's joint traffic-management proposal and said the strait should never return to its pre-war arrangement. The molecule test has therefore moved decisively in the wrong direction: neither safe, insurable passage nor durable normal flows have been demonstrated.

Sources: IEA: Oil Market Report July 2026 S&P Global: Hormuz traffic falls to 11 ships AP: Iran rejects Oman traffic proposal

Update August 19, 2026: same. Compared to July 29: mixed, and the molecule test is still not passed — but the gap between headlines and molecules has never been this measurable. On August 11, US Energy Secretary Chris Wright said the seven-day average leaving Hormuz was almost 9 million barrels per day, plus another 5–7 million barrels per day via bypass routes. Independent trackers cannot reproduce that. Kpler counted 84 vessel transits for the entire week, against more than 100 per day before the war, and estimates roughly 5 million barrels per day by sea; Commodity Context puts it nearer 7 million barrels per day by sea and 4 million through pipelines. Kpler’s Matt Smith said it is “not possible to reconcile the disparity between what we see and what he is quoting”. CNBC reported Hormuz traffic near a three-month low on August 12 with the US–Iran deal in doubt, and Iran still describes the strait as shut. Meanwhile VLCCs are loading west of Hormuz and transferring cargo ship-to-ship in the Gulf of Oman, a workaround that moves barrels without proving the strait itself works. The genuine improvement is in buffers rather than in the strait: EIA’s latest published report shows Cushing rebuilt to 22.6 million barrels for the week ending August 7, up 1.6 million on the week and well above the July 24 low of 18.599 million, with commercial crude up 17.4 million barrels to 424.4 million — although that level is still about 2% below the five-year average — and the SPR at 298.7 million barrels. So inventories are refilling and screen prices fell on the 9-million-barrel claim, but transit counts, insurance, and verified delivery still do not demonstrate normalization. Molecules, not headlines, remains exactly the right filter.

Sources: CNN: Hormuz traffic vs administration claims CNBC: traffic near three-month low Fortune: Iran says strait shut EIA: Weekly Petroleum Status Report EIA: Cushing weekly stocks

Update June 17, 2026: worse. Compared to the original/latest previous text: the checklist matters more. EIA still shows low Cushing stocks for the week ending June 5, while the MOU path still depends on actual signatures, mine risk, insurance, and Lebanon. Watch ships, inventories, product prices, and credit conditions, not press language. EIA: Cushing stocks

Sources: EIA: Weekly Petroleum Status Report EIA: Cushing stocks Guardian: Iran/Lebanon condition

Update June 24, 2026: same. Compared to June 17: the checklist is still the right filter. WPSR shows a large crude draw and Cushing near 20 million barrels, while FRED shows high-yield spreads still calm. That is exactly why both physical data and credit data need to be watched together.

Sources: EIA: Weekly Petroleum Status Report EIA: Cushing stocks FRED: high-yield spread

Update July 1, 2026: same. Compared to June 24: the checklist is still the right filter. The latest published EIA/WPSR data still cover the week ending June 19 with Cushing at 18.96 million barrels, while FRED shows high yield at 2.80% on June 29, the 10-year at 4.38% on June 29, and the 10y-3m spread at 0.57% on June 30. Credit markets therefore still look calmer than the physical inventory picture.

Sources: EIA: Weekly Petroleum Status Report EIA: Cushing weekly stocks FRED: high-yield spread FRED: 10-year Treasury FRED: 10y-3m spread

Update July 29, 2026: worse. Compared to July 1: the checklist is mixed, but the net signal is worse. The latest EIA report shows Cushing at 18.599 million barrels, down 771,000 barrels in the week ending July 24; commercial crude stocks fell by 7.167 million barrels and the SPR by 3.797 million barrels. Shipping through Hormuz has deteriorated again and the 10-year Treasury rose from 4.38% on June 29 to 4.65% on July 27, although high-yield remained comparatively calm at 2.84% on July 28. One genuine improvement is inflation: June CPI fell 0.4% month over month and slowed from 4.2% to 3.5% year over year, but energy was still 15.7% higher than a year earlier. The checklist therefore still warns that easing headline inflation and calm credit spreads do not offset weaker physical buffers, disrupted logistics, and higher long-term rates.

Sources: EIA: Weekly Petroleum Status Report, July 29 EIA: weekly petroleum stocks BLS: CPI June 2026 FRED: high-yield spread FRED: 10-year Treasury S&P Global: Hormuz traffic

Update August 19, 2026: same. Compared to July 29: the checklist is doing exactly what it was built to do — the headline inflation number improved while the household and long-rate legs got worse. Inflation: July CPI rose only 0.1% on the month and eased to 3.4% year over year from 3.5% in June, with core at 2.5%, down 0.1 percentage point. Energy fell 1.5% on the month after a 5.7% drop in June, but is still 14.7% higher than a year earlier, and gasoline is up 24.6% year over year. Inventories: the physical buffer rebuilt. EIA shows Cushing at 22.6 million barrels for the week ending August 7, up 1.6 million on the week and well above the July 24 low of 18.599 million; commercial crude rose 17.4 million barrels to 424.4 million, although that level is still about 2% below the five-year average, and the SPR stood at 298.7 million barrels. Credit: the high-yield spread tightened to 2.71% on August 12 from 2.84% on July 28, and investment-grade markets absorbed a $56 billion issuance week without meaningful spread disruption. Long rates: the opposite direction. The 10-year Treasury reached 4.75% on August 18, its highest level in 19 months, up from 4.65% on July 27, with the 2s/10s curve steepening to about 52 basis points. Real economy: this is the clearest deterioration. The national average gasoline price hit a record seasonal high around $4.03–4.10 per gallon, roughly 30% above a year earlier; July retail sales fell 0.6%, the largest drop since May 2025; and University of Michigan sentiment fell about 8% to 51 in early August from 55.2 in July, while year-ahead inflation expectations rose to 4.3%. So cooling headline inflation and calm credit spreads still do not offset record fuel costs, weakening consumption, falling confidence, and the most expensive long-term funding in 19 months. Reading all the channels together, rather than the CPI print alone, remains the entire point of this list.

Sources: CNBC: CPI July 2026 EIA: Weekly Petroleum Status Report FRED: high-yield spread FRED: 10-year Treasury Bloomberg: 10-year at highest since 2025 AAA: fuel prices August 2026 University of Michigan: Surveys of Consumers

Status June 3, 2026: worse

• Oil/Hormuz: the old point still holds. S&P Global/Platts describes Hormuz traffic as still severely constrained by physical security risk and very expensive insurance, so “open” on paper is not the same as normal commercial throughput. S&P Global/Platts: defining “open”
• Private credit: the stress is clearer than when this menu was written. Reuters reported on May 29 that unrealized losses at US private-credit lenders deepened in the first quarter, while attention remains on valuations, non-accruals, and redemption requests. Reuters/Kitco: private-credit losses deepen
• Japan/the yen: the Japan track is also more acute. Reuters reported on June 3 that the yen was again pushed toward 160 per dollar and Japanese authorities issued fresh warnings, keeping the link between oil, yen stress, and global rates relevant. Reuters/Investing.com: yen near 160

Update June 17, 2026: worse. Compared to June 3: same three weak points, more loaded. Oil/Hormuz is still not physically normalized, private credit remains exposed to markdowns, and Japan/yen risk is still a liquidity channel. The thesis is not new, but the margin for error is smaller than on June 3.

Sources: S&P Global/Platts: Hormuz open definition Reuters/Kitco: private-credit losses Reuters/Investing.com: yen near 160

Update June 24, 2026: worse. Compared to June 17: the same three points remain, but oil is weaker because Cushing fell to 20.03 million barrels. Private credit remains an FSB risk and Japan is still in the post-BOJ-hike regime at 1.0%. This is not a new mechanism, but it leaves less buffer.

Sources: EIA: Cushing stocks FSB: private credit vulnerabilities BOJ: policy decision

Update July 1, 2026: same. Compared to June 24: the same three weak points remain. Cushing is now below the practical 20-million-barrel line, high-yield is only mildly higher at 2.80%, and the 10-year has eased to 4.38%. That is mixed, but it does not change the core mechanism: oil, private credit, and Japan/liquidity remain linked risk channels.

Sources: EIA: Cushing weekly stocks FRED: high-yield spread FRED: 10-year Treasury BOJ: policy decision

Update July 29, 2026: worse. Compared to July 1: all three channels are more clearly stressed. Oil: North Sea Dated fell to about $68 per barrel in early July and then rose to about $77 after fighting resumed, but the IEA says refined-product and LPG exports from the Gulf were still below half their pre-war level and product margins reached four-year highs; S&P Global subsequently recorded only 11 Hormuz transits on July 12. The relative fall in screen crude therefore did not mean normal physical product supply. Private credit: a Reuters analysis found 28 of 53 listed BDCs loss-making, versus 12 a year earlier, as loan values and debt costs worsened. Japan: the BOJ recorded the yen at 163.68-163.70 per dollar at 17:00 JST on July 29, while Japan's 10-year yield had reached a 30-year high of 2.865% earlier in July. A weak yen and higher domestic yields keep both intervention and capital-repatriation/carry risks live.

Sources: IEA: Oil Market Report July 2026 S&P Global: Hormuz traffic Reuters: listed BDC losses BOJ: exchange rates, July 29 Reuters: Japanese bond yields

Update August 19, 2026: worse. Compared to July 29: all three clarifications moved further in the wrong direction, and the Japan leg in particular has stopped being theoretical. Oil: the relative relief of early July is gone. Brent traded around $92.42 per barrel on August 18, a fourth consecutive session of gains and roughly $25.68 higher than a year earlier, with WTI opening at $82.43 on August 17. Prices rose precisely because the US and Iran showed little sign of a deal to reopen Hormuz, so this is scarcity pricing rather than demand strength — the opposite of the falling screen prices that looked reassuring in July. Private credit: the marks are now moving. Aggregate unrealised losses across reviewed BDCs reached 2.35% of net asset value in the first quarter, the deepest quarterly decline since the second quarter of 2022, and in mid-August Blackstone Secured Lending Fund posted its largest quarterly NAV decline in six years, to $25.53 per share, driven by portfolio markdowns rather than a jump in non-performing loans. The spread advantage that justified the asset class has compressed from more than 300 basis points over leveraged loans on 2017–2018 vintages to under 100 basis points by the first quarter of 2026. Some coverage reads the latest results as stabilisation away from the worst case, which is fair — but it is stabilisation at visibly weaker marks. Japan: this is the clearest change. Japan's 10-year yield reached 2.95% on August 18, the highest since 1996, up from the 2.865% recorded in July. After the joint US–Japan intervention in late July, in which both countries bought yen, the currency has strengthened from 163.68–163.70 per dollar on July 29 to roughly 159.7 now, and traders expect a BOJ rate hike as early as September. Higher Japanese yields, a firmer yen, and an imminent hike together form the textbook trigger for yen-carry unwinding, so the liquidity channel described in this section is no longer a hypothetical risk but an active one.

Sources: Fortune: oil price August 18 Bloomberg: Japan yields highest since 1996 Trading Economics: Japan 10-year yield Benzinga: BDC valuation doubts Private Equity Wire: BDC results

Status June 3, 2026: worse

Oil rose again on June 3 as new Middle East hostilities flared and Iran-US talks showed little clear progress. That strengthens, rather than weakens, the August risk because physical energy and logistics remain the stress point. Reuters/Investing.com: oil rises as talks stall

Update June 17, 2026: worse. Compared to June 3: the risk picture is worse. The August window looks more relevant because Cushing remains near the red zone and the MOU/Hormuz setup is still more paper process than proven physical normalization. If flows and inventories do not turn quickly, the lag reaches energy, shipping, prices, and credit.

Sources: EIA: Cushing stocks Guardian: Iran deal claims FSB: private credit vulnerabilities

Update June 24, 2026: worse. Compared to June 17: the August window is more loaded. Cushing is now just above the practical 20-million-barrel line, and the IEA June outlook says recovery in oil flows will not be immediate. That makes the lag into prices, freight, and credit more relevant.

Sources: EIA: Cushing stocks IEA: Oil Market Report June 2026

Update July 1, 2026: worse. Compared to June 24: the August risk is more loaded because the latest official Cushing figure is now 18.96 million barrels, below the practical 20-million-barrel line. FRED does not show open credit panic, but the physical buffer has moved from red-zone warning to acute warning territory.

Sources: EIA: Cushing weekly stocks EIA: Weekly Petroleum Status Report FRED: high-yield spread

Update July 29, 2026: worse. Compared to July 1: the August scenario is no longer only a forecast; parts of it are visible in energy, logistics, and household costs. EIA reports Cushing at 18.599 million barrels for the week ending July 24, with commercial crude down 7.167 million barrels and the SPR down 3.797 million barrels in one week. The absolute US gasoline price was $4.11 per gallon on July 26—up $0.21 from a month earlier and $0.96 from a year earlier—and consumer confidence fell to 90.8 in July from 92.2 in June. This supports the original claim of clear economic pressure by August, but not a claim of broad collapse: June CPI fell 0.4% month over month, although energy remained 15.7% higher year over year; the IMF says the world economy has held up better than feared, and public credit spreads still do not show panic.

Sources: EIA: Weekly Petroleum Status Report, July 29 EIA: weekly petroleum stocks AP: US gasoline prices AP: July consumer confidence BLS: CPI June 2026 IMF: July 2026 outlook FRED: high-yield spread

Update August 19, 2026: worse. Compared to July 29: the August window this section pointed at has now effectively closed, and on the real-economy side the claim largely verified — while markets still refuse to confirm it. Labour market: the clearest turn. BLS reports that nonfarm payrolls fell by 23,000 in July against an expected gain of about 83,000, with government down 53,000 and private payrolls up only 30,000; May was revised down 66,000 and June down 37,000, leaving those two months a combined 103,000 lower than first reported. Average hourly earnings growth slipped to 3.2% year over year, the weakest since May 2021 — below the 3.4% CPI rate, so real wages are now falling. Unemployment was 4.1%. Households: July retail sales fell 0.6%, the largest drop since May 2025; University of Michigan sentiment fell about 8% to 51 in early August from 55.2 in July; and the national average gasoline price reached a record seasonal high near $4.03–4.10 per gallon, roughly 30% above a year earlier. Costs and funding: Brent traded around $92.42 on August 18, some $25.68 above a year earlier, the 10-year Treasury hit 4.75%, a 19-month high, and the 30-year reached its highest level in nearly two decades. So “clear economic problems by August” is now well supported: falling payrolls, negative real wage growth, shrinking retail sales, collapsing confidence, and record fuel costs. What has not happened is the systemic break. The high-yield spread was 2.71% on August 12, investment grade absorbed a $56 billion issuance week without disruption, and the S&P 500 closed at 7,691.76 on August 18 — down 0.7% on the day and falling for a third session on bond yields and oil, but still roughly 3.5% above the July 28 level of 7,428.78. Consensus still puts recession probability near 25% over the next twelve months. Read as a forecast of visible economic pressure by August, the original 80% call holds up well; read as a forecast of a crash, it does not. August delivered a measurably weaker economy, not a systemic event.

Sources: BLS: Employment Situation, July 2026 CNBC: jobs report July 2026 AAA: fuel prices August 2026 University of Michigan: Surveys of Consumers Fortune: oil price August 18 TheStreet: markets August 18 FRED: 10-year Treasury

Status June 3, 2026: worse

Because oil and shipping risk are being repriced upward after fresh Middle East headlines, the EU’s energy sensitivity is more relevant, not less. The stagflation risk is therefore not off the table. Reuters/Investing.com: oil rises as talks stall

Update June 17, 2026: worse. Compared to June 3: unchanged to worse. The EU risk is not solved by a headline deal while physical energy, freight, and insurance remain uncertain. Europe’s problem is its cost base: energy, chemicals, transport, fertilizer, and weaker industrial margins.

Sources: ECB: energy and inflation risk World Bank/Reuters: energy prices IEA: Middle East energy markets

Update June 24, 2026: same. Compared to June 17: the EU risk is still the same type of problem. The IEA sees lower oil demand as high fuel prices and product availability bite, which is weak demand rather than healthy normalization. Europe’s energy, transport, and industrial cost base remains the core issue.

Sources: IEA: Oil Market Report June 2026 IMF: Global Financial Stability Report

Update July 1, 2026: same. Compared to June 24: the EU mechanism is still not solved. Cushing is weaker, but the public credit and rate indicators are mixed rather than panicked. That leaves the same stagflation channel in place: energy, freight, chemicals, fertilizer, and industrial margins remain the relevant tests.

Sources: EIA: Cushing weekly stocks FRED: 10-year Treasury IEA: Oil Market Report June 2026

Update July 29, 2026: same. Compared to July 1: the EU signal remains mixed in exactly the stagflationary way described here. Euro-area inflation eased from 3.2% in May to 2.8% in June and fell 0.1% month over month, which is real relief. But EU motor-fuel prices were still 13.7% higher than a year earlier. On July 23 the ECB kept rates unchanged, said energy prices remained well above pre-conflict levels and warned that the shock's full inflation effect had yet to emerge while high energy costs were restraining growth. This is persistent stagflation risk, not proof that full stagflation has already arrived.

Sources: Eurostat: June inflation Eurostat: June motor-fuel prices ECB: July monetary policy decision ECB: July economy summary

Update August 19, 2026: worse. Compared to July 29: July's relief has reversed. Eurostat shows euro-area annual inflation rising to 2.9% in July from 2.8% in June, driven by energy inflation accelerating to 10.0% from 8.5% as US–Iran hostilities resumed. Core rose to 2.5% from 2.4%, services to 3.3% from 3.2%, and non-energy industrial goods to 0.9% from 0.7%, so the pressure is no longer confined to the energy line. The ECB held its three key rates unchanged on July 23, after raising them by 25 basis points on June 11 to 2.40%, 2.65% and 2.25%. Its own baseline now sees growth of only 0.8% in 2026, 1.2% in 2027 and 1.5% in 2028 — a downward revision for both 2026 and 2027 that the ECB attributes to the war's effect on commodity markets, real incomes and confidence, with upside risks to inflation and downside risks to growth. Reaccelerating prices, broadening core pressure, and downgraded growth together are the textbook definition this section described, and it is now the ECB's own baseline rather than an outside warning.

Sources: Eurostat: euro area annual inflation ECB: monetary policy decisions ECB: Economic Bulletin 5/2026

Status June 3, 2026: worse

Private-credit stress has continued to become more visible. Reuters reported on May 29 that unrealized losses at US private-credit lenders deepened to the worst level since 2022. Reuters/Kitco: private-credit losses deepen

Update June 17, 2026: worse. Compared to June 3: the risk picture is worse. The credit-stress window is more open now because inventories and geopolitics remain pressured while funding costs, private credit, and refinancing remain weak points.

Sources: Reuters: private credit markdowns FSB: private credit vulnerabilities IMF: Global Financial Stability Report

Update June 24, 2026: worse. Compared to June 17: the risk is slightly worse. Cushing is almost at acute levels, the FSB still warns about concurrent vulnerabilities, and AI debt flows are clearer. That is the kind of setup where markets can look calm until several channels are priced at once.

Sources: EIA: Cushing stocks FSB: financial-stability vulnerabilities Axios: AI debt

Update July 1, 2026: same. Compared to June 24: the market/credit setup is still fragile rather than openly broken. Cushing is worse at 18.96 million barrels, while high-yield at 2.80% and the 10-year at 4.38% still do not show broad public-market panic. That keeps the risk conditional but elevated.

Sources: EIA: Cushing weekly stocks FRED: high-yield spread FRED: 10-year Treasury

Update July 29, 2026: worse. Compared to July 1: hidden credit risk has become more visible and long-term funding is more expensive, although public markets are not in crisis. Reuters found 28 of 53 listed BDCs loss-making versus 12 a year earlier. The high-yield spread moved only from 2.80% on June 29 to 2.84% on July 28, but the 10-year Treasury rose from 4.38% to 4.65% on July 27. The S&P 500 was 7,428.78 on July 28, about 0.9% below June 30—not a crash. The year-end stress risk is therefore higher because private-credit losses and refinancing costs are worsening beneath a still-orderly public-market surface.

Sources: Reuters: listed BDC losses FSB: private-credit vulnerabilities FRED: high-yield spread FRED: 10-year Treasury FRED: S&P 500

Update August 19, 2026: same. Compared to July 29: the same split screen, now with a clear repricing at the long end. What still looks calm: the high-yield spread tightened to 2.71% on August 12 from 2.84% on July 28, investment grade absorbed a $56 billion issuance week without meaningful spread disruption, and the S&P 500 closed at 7,691.76 on August 18 — still about 3.5% above its July 28 level of 7,428.78, despite three consecutive down sessions on bond yields and oil. What does not: the 10-year Treasury reached 4.75% on August 18, a 19-month high, the 30-year hit its highest level in nearly two decades, and the 10y–3m spread widened to about 0.96 percentage points — a 10-year at 4.75% against a 3-month bill at 3.79% — from 0.57% at the end of June. That is a bear steepening driven by the long end, not by policy easing. Private credit: aggregate unrealised losses across reviewed BDCs reached 2.35% of net asset value in the first quarter, the deepest quarterly decline since the second quarter of 2022, and Blackstone Secured Lending Fund posted its largest quarterly NAV decline in six years in mid-August. Stress is therefore accumulating where marks are discretionary and funding is long-dated, while the spreads that would signal panic stay tight. This is precisely the configuration the section warned about: the risk of stress before year-end is not lower merely because public markets still look orderly.

Sources: FRED: high-yield spread FRED: 10y–3m spread Bloomberg: bond rout TheStreet: markets August 18 Private Equity Wire: BDC results

Status June 3, 2026: worse

The FSB said on June 1 that elevated sovereign debt, shorter maturities, and more leveraged trading strategies leave bond markets vulnerable to shocks. That fits the view that energy shock plus credit stress can more easily become systemic stress. FSB: new financial-stability vulnerabilities

Update June 17, 2026: worse. Compared to June 3: mixed but more dangerous. This is still not automatic 2008 panic, but the combination of energy, inventories, credit, bonds, and AI debt makes a physical shock more likely to become systemic if markets must price everything at once.

Sources: FSB: financial-stability vulnerabilities IMF: Global Financial Stability Report Axios: AI debt

Update June 24, 2026: same. Compared to June 17: still high but conditional systemic risk. The high-yield spread is not in panic, but Cushing, private credit, long rates, and AI financing remain amplifiers. Systemic stress requires several of them to ignite together.

Sources: FRED: high-yield spread EIA: Cushing stocks FSB: vulnerabilities

Update July 1, 2026: same. Compared to June 24: still high but conditional systemic risk. The physical inventory signal is worse, but public credit spreads are not yet confirming a systemic break. The same combination must be watched: inventories, private credit, long rates, AI financing, and leverage.

Sources: EIA: Cushing weekly stocks FRED: high-yield spread FSB: vulnerabilities

Update July 29, 2026: same. Compared to July 1: latent systemic risk is higher in physical energy and private credit, but a 2008-like break is still not confirmed. On July 24 the St. Louis Fed Financial Stress Index was -0.8263 and the Chicago Fed National Financial Conditions Index was -0.554; negative readings indicate below-average stress and looser-than-average financial conditions. High-yield was 2.84% and the VIX 18.21 on July 28. Against that calm surface stand Cushing below 20 million barrels and newly visible BDC losses. The systemic scenario therefore remains high-impact and conditional, while the original 60–65% probability should still be treated as an aggressive judgment rather than an observed market fact.

Sources: FRED: St. Louis Fed Financial Stress Index FRED: Chicago Fed financial conditions FRED: high-yield spread FRED: VIX FSB: private-credit vulnerabilities

Update August 19, 2026: worse. Compared to July 29: the conditional in this section — systemic stress if Hormuz stays shut — has moved closer to being triggered, because the condition itself is holding. Kpler counted 84 vessel transits through the strait in a week, against more than 100 per day before the war, and independent trackers put seaborne flows at roughly 5–7 million barrels per day against the US administration's claim of almost 9 million. The IEA warned in mid-August that reopening is becoming more pressing because the world is burning through its oil stockpiles at a rapid pace. That is the precise mechanism by which a conditional risk becomes an actual one: buffers are finite, and drawing them down converts a shipping problem into a pricing and credit problem with a lag. Brent at about $92.42 on August 18, some $25.68 above a year earlier, is that lag starting to show. What still has not happened is the financial break — the high-yield spread was 2.71% on August 12 and investment grade absorbed a $56 billion issuance week without disruption. So the physical precondition is intensifying while the financial trigger has not fired. That is a worse setup than in July rather than a safer one, because the buffer that has absorbed the shock so far is measurably smaller than it was.

Sources: CNN: IEA warning on stockpiles CNN: Hormuz traffic vs claims Fortune: oil price August 18 FRED: high-yield spread

Status June 3, 2026: worse

The clean “everything blows over” scenario looks even weaker when oil rises on new hostilities and talks are described as slow. To change that assessment, actual flows and insurance need to normalize, not just headlines. Reuters/Investing.com: oil rises as talks stall

Update June 17, 2026: worse. Compared to June 3: a clean resolution is still unlikely. For this to blow over neatly, the market needs more than an MOU headline: working Hormuz flows, a stable Lebanon track, lower insurance risk, stopped inventory drawdowns, and calm credit. That full combination is still not proven.

Sources: Guardian: Iran deal claims EIA: Cushing stocks FSB: private credit vulnerabilities

Update June 24, 2026: same. Compared to June 17: a clean resolution is still not proven. Better traffic through Hormuz is a real counterweight, but Cushing kept falling and credit/AI debt remain in place. A clean blow-over needs improvement in every channel, not only in the oil price.

Sources: IEA: supply readjustment EIA: Cushing stocks Axios: AI debt

Update July 1, 2026: same. Compared to June 24: a clean resolution is still not proven. The latest official inventory data remain at Cushing 18.96 million barrels, while high-yield and rates look calmer than the physical buffer. The clean outcome still requires several channels to improve together.

Sources: EIA: Cushing weekly stocks EIA: Weekly Petroleum Status Report FRED: high-yield spread

Update July 29, 2026: same. Compared to July 1: a clean resolution remains effectively outside the base case. Renewed fighting reduced visible Hormuz traffic to 11 ships on July 12, Cushing is still below 20 million barrels, and private-credit losses are more visible. Those facts mean some economic damage has already occurred, so “cleanly” is no longer a good description even if escalation ends soon. The counterweight is that June inflation eased and the IMF says the global economy has held up better than feared. A messy but manageable outcome remains plausible; an outcome with no meaningful after-effects does not.

Sources: S&P Global: Hormuz traffic EIA: Weekly Petroleum Status Report Reuters: listed BDC losses BLS: CPI June 2026 IMF: July outlook

Update August 19, 2026: same. Compared to July 29: nothing has moved toward the clean resolution this section priced at 0–1%. Iran still describes the strait as shut, the US and Iran have shown little sign of agreement, and OFAC's Iran-related General License X authorises the transactions it covers only through 12:01 a.m. EDT on August 21, 2026 — so even the partial relief currently in place has a near-term expiry rather than a durable settlement behind it. Treasury has meanwhile issued its eighth sanctions action of 2026 against Iran's alleged shadow-banking network under the “maximum pressure” campaign, which is the opposite of a de-escalation track. A tidy unwind would require simultaneous progress on shipping, insurance, mine clearance, sanctions architecture, and Lebanon, and none of those has been demonstrated. The 0–1% estimate therefore still looks right, and for the original reason: everything blowing over neatly is not one event but a stack of them, and stacks do not clear at once.

Sources: OFAC: Iran-related General License State Department: Iran sanctions Fortune: Iran says strait shut

Status June 3, 2026: worse

The August window looks more relevant, not less. The latest Reuters data shows deeper unrealized losses and still-high non-cash interest income in private credit. Reuters/Kitco: private-credit losses deepen

Update June 17, 2026: worse. Compared to June 3: worse. Credit stress before August remains a reasonable main window. Energy and inventory stress first hit margins, then refinancing, markdowns, and tighter lending conditions.

Sources: Reuters/Kitco: private-credit losses FSB: private credit vulnerabilities FRED: high-yield spread

Update June 24, 2026: same. Compared to June 17: the credit-stress window remains relevant. Public high-yield spreads are still calm, but the FSB private-credit warning and higher physical-cost risk mean stress can arrive through markdowns and refinancing before index spreads scream.

Sources: FRED: high-yield spread FSB: private credit vulnerabilities

Update July 1, 2026: same. Compared to June 24: the credit-stress window is still relevant. High-yield has moved to 2.80%, up from June 22 but still not panic. That means the visible spread market is calm enough to be a counterweight, while private-credit and refinancing risk remain the likely earlier channels.

Sources: FRED: high-yield spread FSB: private credit vulnerabilities FRED: 10-year Treasury

Update July 29, 2026: worse. Compared to July 1: the pre-August window has produced clearer private-credit stress, but not a public credit panic. Reuters' standardized analysis shows 28 of 53 listed BDCs loss-making in the first quarter, with average profit across the group falling from $26 million a year earlier to a $7.6 million loss. Meanwhile high-yield was still only 2.84% on July 28, while the 10-year Treasury had risen to 4.65% on July 27. The sequence described here is therefore visible: markdowns and refinancing pressure are appearing before broad index spreads break.

Sources: Reuters: listed BDC losses FSB: private-credit vulnerabilities FRED: high-yield spread FRED: 10-year Treasury

Update August 19, 2026: same. Compared to July 29: August has arrived, and the answer to this section's own question — how fast credit stress becomes visible — turns out to be “slowly, and not where the headlines look”. Public credit did not crack. The high-yield spread was 2.71% on August 12, tighter than the 2.84% recorded on July 28, and investment grade absorbed a $56 billion issuance week without meaningful spread disruption. The stress appeared instead in the two slower channels this section identified. Private marks: aggregate unrealised losses across reviewed BDCs reached 2.35% of net asset value in the first quarter, the deepest quarterly decline since the second quarter of 2022, and Blackstone Secured Lending Fund recorded its largest quarterly NAV decline in six years in mid-August. Households: the mortgage delinquency rate was 4.37% in the second quarter — down 7 basis points on the quarter but up 44 basis points from a year earlier — and the pattern is strikingly uneven, with FHA serious delinquencies up 227 basis points year over year against just 6 basis points for conventional loans and 31 for VA. Credit card, student loan and auto performance continued to deteriorate. The timing question is therefore largely answered: discretionary marks and lower-income borrowers move first, tradable spreads move last, and that gap is precisely why the public data looked reassuring throughout the window this section was written about.

Sources: FRED: high-yield spread Private Equity Wire: BDC results HousingWire: Q2 mortgage delinquencies National Mortgage News: consumer stress

Status June 3, 2026: worse

The China track looks more relevant. Reuters reported on May 20 that China defends its rare-earth controls as lawful while saying it can cooperate on “reasonable” US concerns, meaning the bottleneck remains an active bargaining tool. Reuters/Investing.com: China rare-earth controls

Update June 17, 2026: same. Compared to June 3: still important. China is not today’s main headline, but rare earths, energy purchases, sanctions, and industrial capacity remain strategic levers. China does not need to be the loudest story to affect the outcome.

Sources: Reuters/EIA: strategic oil inventories Reuters: China stockpiling Reuters/Investing.com: rare-earth controls

Update June 24, 2026: same. Compared to June 17: still strategically important. The IEA describes how replacement flows and inventories helped cushion the Hormuz gap, but that makes the China track more important, not irrelevant: energy, inventories, sanctions, and critical materials remain bargaining power.

Sources: IEA: supply readjustment IEA: Oil Market Report June 2026

Update July 1, 2026: same. Compared to June 24: the China track is still a strategic lever rather than the loudest daily signal. The inventory picture keeps energy bargaining power relevant, and critical materials, sanctions, and oil stockpiles remain part of the same asymmetry.

Sources: EIA: Cushing weekly stocks IEA: Oil Market Report June 2026 Reuters: China stockpiling

Update July 29, 2026: worse. Compared to July 1: China's oil buffer has been used more heavily, while its critical-material leverage has become more explicit. The IEA estimates that Chinese crude stocks fell by 41 million barrels in June, and Reuters reports that June crude imports fell 41.3% to their lowest level in almost a decade. That weakens the inventory cushion, but does not remove the strategic asymmetry: the IEA warns that full implementation of China's rare-earth restrictions could put $6.5 trillion of non-Chinese downstream production at risk, and China added 14 European entities to an export-control list on July 24 in retaliation for EU sanctions. The net change is worse because energy resilience has been tested while export controls remain an active bargaining and retaliation tool.

Sources: IEA: Oil Market Report July 2026 Reuters: China's oil imports Reuters/IEA: rare-earth exposure AP: export controls on EU entities

Update August 19, 2026: worse. Compared to July 29: China has been widening this lever rather than easing it. Beijing placed 10 US companies under new rare-earth export restrictions on June 22 and blocked dual-use shipments to 14 EU firms by July 24. The underlying concentration is unchanged and decisive: China controls roughly 90% of rare-earth processing, about 80% of tungsten and 60% of antimony. The effect now shows up in prices and paperwork rather than in rhetoric — the NdPr alloy benchmark rose 21.4% in a single month through July 1, the 2025–2026 control regime has produced price spikes of up to sixfold, and licence approval rates for European firms have run below 25%. The same lever is being applied to agriculture: China extended restrictions on phosphate fertiliser exports, including DAP, MAP and selected NPK blends, through August 2026, now covering an estimated 50–80% of its export volumes. That matters because it stacks with the Hormuz-driven sulfur and ammonia disruptions covered further down this page — the same buyer is squeezed from two directions at once. The China card described here is therefore no longer a latent option; it is being played incrementally, in exactly the input markets where substitution is slowest.

Sources: Morgan Lewis: China export control enforcement Rare Earth Exchanges: 2026 export controls CSIS: rare earth restrictions one year later Fertilizer prices weekly update

Status June 3, 2026: worse

Valuation risk has not eased. Market commentary on June 2 still described new S&P 500 records, making an already extreme Buffett/valuation picture more stretched rather than cheaper. FXEmpire: S&P 500 at record highs

Update June 17, 2026: worse. Compared to June 3: still an expensive risk picture. This indicator is not a crash timer, but it says the market has little valuation buffer if earnings, rates, or credit conditions start moving the wrong way.

Sources: GuruFocus: Buffett Indicator FRED: US GDP FXEmpire: AI-led rally

Update June 24, 2026: same. Compared to June 17: valuation buffer is still thin. FRED shows the 10-year around 4.51% on June 22, keeping the discount rate high while S&P 500 concentration was already extreme. This is not a timer, but the margin is small.

Sources: FRED: 10-year Treasury S&P 500 factsheet

Update July 1, 2026: worse. Compared to June 24: the valuation buffer is thinner. FRED shows the S&P 500 at 7,499.36 on June 30, above the June 24 level, while physical inventory stress is still unresolved. The 10-year easing helps, but higher equity prices leave less room for disappointment.

Sources: FRED: S&P 500 FRED: 10-year Treasury EIA: Cushing weekly stocks

Update July 29, 2026: better. Compared to July 1: valuation pressure has eased only marginally. The S&P 500 closed at 7,428.78 on July 28, about 0.9% below its June 30 close of 7,499.36. That is a small relative improvement, not a cheap market: GuruFocus put the total-market-cap-to-GDP ratio at 234.3% on July 23, versus 227% cited here in April, while first-quarter nominal GDP was $31.866 trillion. The indicator remains a valuation-risk gauge, not a crash timer.

Sources: FRED: S&P 500 GuruFocus: Buffett Indicator FRED/BEA: nominal GDP

Update August 19, 2026: worse. Compared to July 29: the ratio has risen further, because equities gained while the real economy did not. Readings in mid-August cluster around 230% to 243% of GDP — GuruFocus put the US total market cap to GDP ratio at 229.9% on August 8 and about 240% on August 18, while buffettindicator.org showed 243.4% on August 13. Those numbers should be read as a range rather than a single figure: providers differ on which market aggregate and which GDP vintage they use, and one market-close estimate ran as high as 292%. Even at the low end of the range, the reading sits far above the roughly 166% long-term average and the 165% level commonly treated as significant overvaluation, and multiples above the 81.3% historical median. The direction is what matters most here: the S&P 500 closed at 7,691.76 on August 18, about 3.5% above its July 28 level, in the same period that payrolls fell by 23,000, real wages turned negative, and retail sales dropped 0.6%. A valuation ratio rises either because the numerator climbs or the denominator stalls; right now both are pushing the same way.

Sources: GuruFocus: market cap to GDP Buffett Indicator: live ratio Current Market Valuation: Buffett indicator model TheStreet: markets August 18

Status June 3, 2026: worse

The spread signal remains treacherous: no full panic, but the FSB warned on June 1 that sovereign debt, shorter maturities, and leverage in bond markets leave rates vulnerable. That makes low spreads look more like fragile calm than safety. FSB: vulnerabilities in bond markets

Update June 17, 2026: same. Compared to June 3: calm on the surface, fragile underneath. High-yield spreads are most dangerous when they look calm while physical and credit weaknesses build underneath. This remains a key indicator for a fast turn.

Sources: FRED: high-yield spread FSB: leveraged trading risks

Update June 24, 2026: same. Compared to June 17: the spread is still calm on the surface. FRED shows 2.65% on June 22, below many crisis levels. That is a counterweight, but not a clean bill of health when private credit and physical energy can stress the system outside the daily spread series.

Sources: FRED: high-yield spread FSB: private credit vulnerabilities

Update July 1, 2026: same. Compared to June 24: the spread has moved up a little but is still not panic. FRED shows 2.80% on June 29, up from 2.65% on June 22. That is a mild surface deterioration, but the level still says “fragile calm” rather than open credit stress.

Sources: FRED: high-yield spread FSB: private credit vulnerabilities

Update July 29, 2026: same. Compared to July 1: the spread has widened by only 0.04 percentage point, from 2.80% on June 29 to 2.84% on July 28. The relative move is mildly worse, but the absolute level still signals calm rather than broad high-yield panic. That is an important counterweight; it does not cancel the separate evidence of losses in listed BDCs and private-credit vulnerability.

Sources: FRED: high-yield spread Reuters: listed BDC losses FSB: private-credit vulnerabilities

Update August 19, 2026: better. Compared to July 29: this specific indicator improved. The ICE BofA US high-yield option-adjusted spread was 2.71% on August 12, down from 2.84% on July 28 and from 2.80% at the end of June. Junk credit is therefore being priced as less risky than a month ago, not more. Investment grade told the same story: the primary market absorbed a $56 billion issuance week without meaningful spread disruption. It is worth stating plainly that this cuts against most of the rest of this page. The honest reading is that the high-yield spread measures what tradable public credit believes, and public credit currently believes conditions are fine. The only caution is that the same weeks produced record seasonal fuel prices, a 23,000 fall in payrolls, negative real wage growth, the deepest quarterly BDC markdown since the second quarter of 2022, and a 19-month high in the 10-year Treasury. A tight spread alongside deteriorating fundamentals is either a correct signal that the stress stays contained, or the ordinary pattern of spreads being the last thing to move. This section cannot settle which of those it is; it can only record that the indicator itself is currently benign.

Sources: FRED: ICE BofA high-yield OAS Trading Economics: high-yield OAS Nuveen: fixed income weekly

Status June 3, 2026: worse

The curve is not the acute warning light, but the rates market looks more vulnerable when the FSB is simultaneously pointing to high sovereign debt, shorter maturities, and leverage. Overall, that makes the situation worse even though this specific curve is not classically inverted. FSB: sovereign debt and leveraged trading risks

Update June 17, 2026: same. Compared to June 3: mixed signal. The curve is not the single acute warning light, but rates remain central because high financing costs make every energy and credit shock harder to absorb.

Sources: FRED: 10y-3m spread FSB: bond-market vulnerabilities

Update June 24, 2026: same. Compared to June 17: the mixed signal remains. FRED shows the 10y-3m spread around 0.65% on June 23. The curve is not screaming recession, but it also does not say financing conditions are easy while the 10-year remains high.

Sources: FRED: 10y-3m spread FRED: 10-year Treasury

Update July 1, 2026: same. Compared to June 24: the mixed signal remains. FRED shows the 10y-3m spread at 0.57% on June 30, slightly lower than June 23. The curve is still positive, but financing conditions are not easy when they have to absorb energy, inventory, and credit risk at the same time.

Sources: FRED: 10y-3m spread FRED: 10-year Treasury

Update July 29, 2026: same. Compared to July 1: the 10y-3m spread has risen from 0.57% on June 30 to 0.71% on July 28. A more positive curve is better in relative recession-signal terms, but the absolute financing picture is not easier: the 10-year itself rose to 4.65% on July 27. The curve therefore remains a mixed signal—less recession-like in shape, but with higher long-term borrowing costs.

Sources: FRED: 10y-3m spread FRED: 10-year Treasury

Update August 19, 2026: worse. Compared to July 29: the curve has steepened, but for the wrong reason. The 10-year Treasury reached 4.75% on August 18 while the 3-month bill held at 3.79%, putting the 10y–3m spread near 0.96 percentage points, against 0.57% at the end of June. A steepening curve is normally read as a recovery signal, but the mechanism decides the meaning: this move came from the long end rising, not from the front end falling on rate cuts. The 2s/10s curve widened to about 52 basis points, its steepest since May, and the 30-year reached its highest level in nearly two decades. Bloomberg attributed the move to heavy corporate supply and fears of entrenched inflation rather than to growth optimism. A bear steepening of this kind raises the cost of precisely the long-dated funding this page keeps returning to — the refinancing wall, commercial real estate, and the AI capex build — so the signal is not the benign one the curve shape alone would suggest.

Sources: FRED: 10y–3m spread Trading Economics: 3-month bill yield Bloomberg: bond rout Federal Reserve: H.15 selected interest rates

Status June 3, 2026: worse

The 10-year remains an amplifier of the risk picture. State Street described on June 1 how rising yields continue reshaping the relationship between equities, credit, and Treasuries. State Street: rising yields reshape markets

Update June 17, 2026: same. Compared to June 3: still a pressure point. The 10-year yield remains an amplifier: it affects mortgages, corporate finance, equity valuations, and government interest costs. High or sticky long rates make the refinancing wall harder.

Sources: FRED: 10-year Treasury State Street: rising yields

Update June 24, 2026: same. Compared to June 17: still a pressure point. FRED shows the 10-year at 4.51% on June 22. That is not bond-market panic, but it is high enough to keep mortgages, corporate debt, real estate, and government interest costs under pressure.

Sources: FRED: 10-year Treasury FSB: bond-market vulnerabilities

Update July 1, 2026: better. Compared to June 24: the 10-year has eased a little. FRED shows 4.38% on June 29, down from 4.51% on June 22. That helps discount rates and refinancing at the margin, but the level is still high enough to keep debt, real estate, and weaker credit under pressure.

Sources: FRED: 10-year Treasury FRED: 10y-3m spread

Update July 29, 2026: worse. Compared to July 1: the 10-year has risen by 0.27 percentage point, from 4.38% on June 29 to 4.65% on July 27. That relative increase reverses the small relief recorded on July 1. The absolute level is not a Treasury-market panic, but it is high enough to intensify pressure on mortgages, corporate refinancing, real estate, government interest costs, and equity discount rates.

Sources: FRED: 10-year Treasury FSB: bond-market vulnerabilities

Update August 19, 2026: worse. Compared to July 29: the 10-year moved from 4.65% on July 27 to 4.75% on August 18, its highest level in 19 months, after touching 4.72% on August 17. The 30-year reached its highest in nearly two decades. The level matters more than the move for everything else on this page. The drivers were a heavy corporate supply calendar — including an estimated $1.5 trillion of bond issuance tied to AI companies over the coming years — and concern that inflation stays entrenched, both of which lift term premia regardless of what the Fed does at the front end. What makes this notable is that the rise happened despite benign July CPI and PPI prints and a 0.6% fall in retail sales: weak activity data did not pull long yields down, which is the opposite of the usual reflex. Equity markets registered it, with the S&P 500 falling for a third consecutive session into August 18. Higher long rates alongside a softening economy is the least comfortable combination available for refinancing, housing, and any capital programme financed on long paper.

Sources: FRED: 10-year Treasury Bloomberg: 10-year highest since 2025 TheStreet: 30-year yield and markets Nuveen: fixed income weekly

Status June 3, 2026: worse

The paper price still does not tell the whole story. Reuters reports oil rising on fresh hostilities and slow talks, while S&P Global/Platts warns that a formally “open” Hormuz does not mean normal traffic, insurance, and throughput. Reuters/Investing.com: oil rises S&P Global/Platts: defining “open”

Update June 17, 2026: same. Compared to June 3: paper price is not enough. MOU headlines can push futures around, but the physical cost is set by freight, insurance, delays, product shortages, and whether ships actually move. This point remains central.

Sources: S&P Global/Platts: Hormuz open definition IEA: Oil Market Report Reuters/Investing.com: shipping costs

Update June 24, 2026: same. Compared to June 17: the paper price looks better, but the point stands. The IEA describes supply readjustment underway, but also says normalization takes time. If lower futures arrive while inventories keep falling, that is not a finished physical solution.

Sources: IEA: supply readjustment EIA: Cushing stocks

Update July 1, 2026: worse. Compared to June 24: the paper-versus-physical point is stronger, not calmer. The latest EIA spot-price series available through FRED has WTI at $78.94 on June 22, which is still an elevated oil price, and EIA shows Cushing down to 18.96 million barrels for the week ending June 19. A slightly lower screen price is therefore not physical normalization; it is still high price plus weaker inventory cover.

Sources: FRED/EIA: WTI spot price EIA: Cushing weekly stocks EIA: Weekly Petroleum Status Report

Update July 29, 2026: worse. Compared to July 1: both the relative move and the absolute physical backdrop are worse. WTI was $84.25 on July 27, about 6.7% above the $78.94 level used in the July 1 update. Brent reached $105.32 on July 23 and then fell to $91.82 on July 27; that decline from the monthly spike is relief in the paper price, but $91.82 remains an elevated absolute level. Meanwhile Cushing fell to 18.599 million barrels in the week ending July 24, and the IEA says product margins reached four-year highs as Gulf product and LPG exports remained below half of pre-war levels. Screen-price relief therefore still does not equal normal port availability, freight, or product supply.

Sources: FRED/EIA: WTI spot price FRED/EIA: Brent spot price EIA: July 29 weekly petroleum report IEA: Oil Market Report July 2026

Update August 19, 2026: worse. Compared to July 29: the gap this section describes has closed from the wrong side. Screen prices rose to meet the physical stress instead of physical conditions easing to meet cheap paper. Brent traded around $92.42 on August 18, a fourth consecutive session of gains and about $25.68 above a year earlier, with WTI opening at $82.43 on August 17. In July this section could still note that falling paper prices were masking tight physical product supply; now both are elevated at once, which removes the last comfort available in the earlier reading. The physical side has not improved to justify the move: independent trackers put seaborne flows through Hormuz at roughly 5–7 million barrels per day against an official claim of almost 9 million, Kpler counted 84 transits in a week against more than 100 per day before the war, and VLCCs are loading west of the strait and transferring cargo ship-to-ship in the Gulf of Oman rather than sailing through it. Downstream the divergence is starker still: sulfur prices are up 261.63% year over year, with roughly half the world's traded sulfur normally moving through Hormuz. Screen crude remains the least informative number in the whole chain.

Sources: Fortune: oil price August 18 CNN: Hormuz flows vs claims Lloyd's List: Strait of Hormuz brief Acres U.S.A.: sulfur shortage

Status June 3, 2026: worse

The rally does not improve the risk picture. The market is still described as record-strong and AI-led, which means concentration and valuation risk have increased rather than eased. FXEmpire: AI hardware leads S&P 500 rally

Update June 17, 2026: worse. Compared to June 3: euphoria can continue, but risk is higher. A strong stock market can coexist with a weaker physical economy. That does not make equities irrelevant, but indexes cannot be used as proof that energy, inventories, and credit are healthy.

Sources: S&P 500 factsheet S&P 500 Equal Weight Reuters/Yahoo Finance: AI fervor

Update June 24, 2026: worse. Compared to June 17: the euphoria is more debt-dependent. The AI rally is supported not just by earnings dreams but by larger bond and financing flows. That makes equity strength more vulnerable if rates or credit conditions start to bite.

Sources: Axios: AI debt boom S&P 500 factsheet

Update July 1, 2026: worse. Compared to June 24: the divergence is larger. FRED shows the S&P 500 at 7,499.36 on June 30, above the June 24 close, while Cushing remains below the acute 20-million-barrel line. Equity strength therefore still does not disprove the physical and credit-risk thesis.

Sources: FRED: S&P 500 EIA: Cushing weekly stocks FRED: high-yield spread

Update July 29, 2026: better. Compared to July 1: the equity-versus-physical divergence has narrowed slightly rather than widened. The S&P 500 closed at 7,428.78 on July 28, about 0.9% below June 30, although it was still up 8.5% for the year. That is a modest relative improvement in the euphoria risk, not proof that underlying risks have disappeared: Cushing remained below 20 million barrels and high-yield spreads remained calm at 2.84%. The lesson of this section still holds, but July's index performance is not evidence of a further rocket-like acceleration.

Sources: FRED: S&P 500 AP: July 28 US indexes EIA: Cushing stocks FRED: high-yield spread

Update August 19, 2026: worse. Compared to July 29: the disconnect widened before it began to narrow. The S&P 500 closed at 7,691.76 on August 18, roughly 3.5% above its July 28 level of 7,428.78 — during the same weeks in which payrolls fell by 23,000, May and June were revised down by a combined 103,000, real wages turned negative with earnings growth at 3.2% against 3.4% CPI, retail sales dropped 0.6%, and consumer sentiment fell about 8% to 51. Valuation followed the index: market cap to GDP readings clustered around 230–243%. The last few sessions have finally shown some transmission — August 18 marked a third consecutive decline, driven by a chip selloff, elevated oil, and a 30-year Treasury yield at its highest in nearly two decades. That is the mechanism this section always pointed to. Equities do not reprice directly because the physical economy became more expensive; they reprice when the discount rate moves. The rocket has not turned around, but for the first time in this cycle the bond market is arguing with it rather than financing it.

Sources: TheStreet: markets August 18 CNBC: third day of declines BLS: Employment Situation, July 2026 Buffett Indicator: live ratio

Status June 3, 2026: worse

The warnings now have more support from data points: the AI rally continues while private-credit losses and energy/geopolitical risks have become clearer. The combination therefore looks more charged than it did in mid-May. Reuters/Kitco: private-credit losses deepen FXEmpire: AI rally

Update June 17, 2026: worse. Compared to June 3: the warnings have more support. New AI-debt news and continued energy/MOU uncertainty fit the combined warning: AI bubble risk, private credit, rates, and geopolitical energy stress inside the same system.

Sources: Reuters/Yahoo Finance: Dalio AI bubble Bloomberg: Taleb warning Yahoo Finance: Steve Eisman on private credit

Update June 24, 2026: worse. Compared to June 17: the warning combination has more support. Nvidia bonds, off-balance-sheet-like AI financing, and the FSB private-credit warning point in the same direction: tech growth can be real and still pull more credit risk into the system.

Sources: Axios: Nvidia bond sale Axios: Anthropic compute financing FSB: private credit vulnerabilities

Update July 1, 2026: same. Compared to June 24: the warning combination is unchanged. AI financing, private credit, high valuations, and physical energy stress still point to the same system-level risk, while public high-yield spreads have not yet confirmed a broad break.

Sources: Axios: Nvidia bond sale FSB: private credit vulnerabilities FRED: S&P 500

Update July 29, 2026: worse. Compared to July 1: the warning combination has gained more support from observable data, not merely from prominent names. Reuters' standardized review found 28 of 53 listed BDCs loss-making in the first quarter; the 10-year Treasury rose to 4.65%; and Cushing fell to 18.599 million barrels. The counterevidence remains important: high-yield spreads were still only 2.84% and the S&P 500 was only modestly below June 30, so these facts support a vulnerability warning, not a claim that a broad crash is already underway.

Sources: Reuters: listed BDC losses FRED: 10-year Treasury EIA: Cushing stocks FRED: high-yield spread

Update August 19, 2026: worse. Compared to July 29: rather than track individual commentators, whose statements are hard to verify consistently, this update records what institutions with published mandates have said — a more checkable standard for the same question. The IEA warned in mid-August that reopening Hormuz is becoming more pressing because the world is burning through its oil stockpiles at a rapid pace. The ECB held rates on July 23 while cutting its growth baseline to 0.8% for 2026 and warning that the energy shock's full inflation effect had yet to emerge. The BIS published analysis on the on- and off-balance-sheet borrowing financing the AI infrastructure boom. The FSB and IMF warnings on private credit vulnerabilities cited earlier on this page have since been matched by hard numbers: aggregate BDC unrealised losses at 2.35% of net asset value, the deepest quarterly decline since the second quarter of 2022. The pattern worth noting is not that any single figure is alarming, but that official bodies now describe the same chain this page describes — physical energy, inflation persistence, leveraged financing, private-credit marks — inside their own baselines rather than as tail risks. That is a stronger form of corroboration than a louder warning from any individual.

Sources: CNN: IEA warning on stockpiles ECB: Economic Bulletin 5/2026 BIS: financing the AI infrastructure boom FSB: private credit vulnerabilities

Status June 3, 2026: worse

The refinancing risk is more concrete. Trepp describes June 2026 as a month with clear CMBS refinancing friction where weak cash flow, impaired valuations, and limited refinancing options converge at maturity. Trepp: June 2026 CMBS hard maturities

Update June 17, 2026: same. Compared to June 3: still vulnerable. The refinancing wall becomes more dangerous when cash flows are squeezed by energy and market rates do not provide relief. It does not require panic, only worse terms at maturity.

Sources: Trepp: June 2026 CMBS maturities IMF: Global Financial Stability Report

Update June 24, 2026: same. Compared to June 17: still vulnerable. A 10-year yield around 4.5% means old cheap debt still has to be replaced in an expensive environment. Sticky high rates and weaker cash flows are enough for the wall to do damage.

Sources: FRED: 10-year Treasury IMF: Global Financial Stability Report

Update July 1, 2026: better. Compared to June 24: slightly better at the margin because the 10-year eased to 4.38% on June 29 from 4.51% on June 22. That helps refinancing math a little, but the wall is still there and cash flows can still be pressured by energy, CRE, and private-credit stress.

Sources: FRED: 10-year Treasury IMF: Global Financial Stability Report FSB: private credit vulnerabilities

Update July 29, 2026: worse. Compared to July 1: the small rate relief that supported the previous “better” assessment has been fully reversed. The 10-year Treasury rose from 4.38% on June 29 to 4.65% on July 27, a 27-basis-point deterioration in the benchmark refinancing backdrop. Reuters' finding that 28 of 53 listed BDCs were loss-making adds evidence that weaker borrowers and lenders are already absorbing pressure. This is still a rollover and cash-flow problem rather than a confirmed market-wide funding freeze.

Sources: FRED: 10-year Treasury Reuters: listed BDC losses IMF: Global Financial Stability Report

Update August 19, 2026: worse. Compared to July 29: the wall itself is unchanged, but the price of climbing it has risen to a multi-decade high. The 10-year Treasury hit 4.75% on August 18, a 19-month high, and the 30-year reached its highest level in nearly two decades — and long rates, not policy rates, are what refinancing is priced against. The maturities remain enormous: roughly $1.0 trillion of US commercial real estate loans matured in 2025, with well over $1.5 trillion reaching maturity by the end of 2026 and some estimates putting 2026 alone as high as $1.8 trillion, while about $1.2 trillion of leveraged debt needs to find a home over the next few years. There is a real counterpoint, and it deserves stating: agencies expect the US leveraged loan default rate to fall to about 3.0% by October 2026 from 5.3% a year earlier, with Europe improving to 2.4% from 3.8%, helped by strong refinancing activity that has pulled 2028–2029 maturities forward. But the quality beneath that has thinned. Average interest coverage on the US leveraged loan index is around 4.6x, well below the roughly 6x of 2022, and floating-rate instruments accounted for 59% of everything that defaulted in 2024 despite representing only about half of outstanding leverage. Lower projected defaults, at materially higher long rates and thinner coverage, describe a wall that has been postponed rather than dismantled.

Sources: PitchBook: 2026 distressed credit outlook Moody's: leveraged finance and CLOs 2026 MMG: 2026 CRE refinancing wall FRED: 10-year Treasury

Status June 3, 2026: worse

The yen track is more acute. Reuters reported on June 3 that the yen was pushed toward 160 per dollar and Japanese authorities warned again, making the carry/liquidity risk more immediate. Reuters/Investing.com: yen near 160

Update June 17, 2026: worse. Compared to June 3: worse. The Bank of Japan raised the policy rate to 1.0% on June 16, the highest level since 1995, roughly 31 years ago. Officially, the reason is inflation risk: broader price pressure, rising inflation expectations, and the risk that underlying CPI moves above the 2% target. In practice, the weak yen matters too, because a currency near 160 per dollar makes imported energy, food, and raw materials more expensive, which can turn into imported inflation and political pressure from households. That makes the yen-carry/liquidity channel more dangerous than the June 3 snapshot. Higher Japanese rates mean cheap-yen funding is less stable, and crowded positions can unwind faster if energy, rates, and credit stress hit together.

Sources: BOJ: June 16 policy decision Reuters/Investing.com: yen near 160 BIS Annual Economic Report 2025

Update June 24, 2026: same. Compared to June 17: same risk channel. The BOJ hike to 1.0% remains a new regime shift, making cheap yen funding less automatic. No new panic is required for carry positions to have less margin for error.

Sources: BOJ: policy decision FRED: 10-year Treasury

Update July 1, 2026: same. Compared to June 24: the Japan/carry channel remains the same risk. The BOJ hike is still the regime change, and lower US long rates help only at the margin. If energy, credit, or equity stress hits, crowded cheap-yen positions still have less margin for error than before.

Sources: BOJ: policy decision FRED: 10-year Treasury BIS Annual Economic Report 2025

Update July 29, 2026: worse. Compared to July 1: both the currency and bond sides of the carry channel deteriorated. The Bank of Japan recorded ¥163.68–163.70 per dollar at 17:00 JST on July 29, versus ¥161.82–161.83 on June 29—a roughly 1.2% weakening of the yen. Japan Bond Trading data show the 10-year JGB yield reached 2.865% on July 8 and was still 2.775% on July 28. A weaker yen raises imported-cost pressure while higher domestic yields make cheap-yen funding less stable; that combination reduces the margin for leveraged carry positions, although no broad forced unwind is yet visible.

Sources: Bank of Japan: July 29 exchange rates Bank of Japan: June 29 exchange rates Japan Bond Trading: JGB historical rates BIS: yen-carry spillovers

Update August 19, 2026: worse. Compared to July 29: this is the section where the risk stopped being a scenario. Japan's 10-year yield reached 2.95% on August 18, the highest since 1996, after rising as much as 5.5 basis points to 2.93% — up from the 2.865% recorded in July. Bonds fell on fiscal concerns and growing speculation that the Bank of Japan raises rates as soon as September, after a growing number of policymakers called for a stronger response to inflationary pressure. The currency leg moved as well: following the joint US–Japan intervention in late July, in which both countries bought yen, the rate has gone from 163.68–163.70 per dollar on July 29 to roughly 159.7, a strengthening of about 2.4%. Each of those three developments on its own reduces the return to borrowing yen in order to fund assets abroad; together they are the classic unwind trigger. The thing to watch is not the yen level but the direction of travel, because a carry trade is unwound by yen appreciation and narrowing differentials — both now underway, while US long rates rise for their own domestic reasons. Repatriation pressure out of Japan meeting a heavy US issuance calendar is precisely the liquidity channel this section described, and it is now running rather than pending.

Sources: Bloomberg: Japan 10-year highest since 1996 Trading Economics: Japan 10-year yield Trading Economics: Japanese yen Bank of Japan

Status June 3, 2026: worse

Real estate and households look weaker. The MBA reported that commercial mortgage delinquencies rose in Q1 2026, and New York Fed/KPMG data shows serious credit-card stress still near financial-crisis levels. MBA: CRE delinquencies increased KPMG/New York Fed: household credit stress

Update June 17, 2026: worse. Compared to June 3: the risk picture is worse. Real estate and households are where the slow hit becomes concrete. Higher costs, credit cards, auto loans, insurance, and refinancing can grow stress even without a dramatic equity-market day.

Sources: MBA: CRE delinquencies KPMG/New York Fed: household credit stress New York Fed: Household Debt and Credit

Update June 24, 2026: same. Compared to June 17: still pressured. There is no new relief in the core mechanism: high rates, higher daily costs, commercial real estate, and household credit. Households and CRE remain where the slow hit becomes measurable.

Sources: New York Fed: Household Debt and Credit MBA: CRE delinquencies FRED: 10-year Treasury

Update July 1, 2026: same. Compared to June 24: still pressured. The 10-year easing to 4.38% helps at the margin, but there is no new broad relief in household debt, CRE refinancing, insurance, food, or energy costs. This remains the slow real-economy transmission channel.

Sources: FRED: 10-year Treasury New York Fed: Household Debt and Credit MBA: CRE delinquencies

Update July 29, 2026: worse. Compared to July 1: household and housing costs have moved in the wrong direction. Freddie Mac's 30-year mortgage rate reached 6.58% on July 23, its highest level in nearly a year. Regular gasoline averaged $4.11 per gallon on July 26, up from $3.90 a month earlier, and the Conference Board's consumer-confidence index fell from 92.2 in June to 90.8 in July. A counterweight is that New York Fed first-quarter data showed small declines in transitions into early mortgage and credit-card delinquency. Overall, current prices and financing costs worsened even though the latest broad delinquency data do not show an abrupt household break.

Sources: Freddie Mac: mortgage rates AP/AAA: gasoline prices AP/Conference Board: consumer confidence New York Fed: Q1 household debt and credit

Update August 19, 2026: worse. Compared to July 29: the slow smash this section described is now measurable, and it is concentrated exactly where the section said it would be — at the bottom of the borrower distribution. The 30-year fixed mortgage rose every week through July to 6.66%, and with the 10-year Treasury at a 19-month high the direction of pressure is unchanged. The headline mortgage delinquency rate looks calm at 4.37% in the second quarter, down 7 basis points on the quarter — but it is up 44 basis points year over year, and the composition is the real story: FHA serious delinquencies rose 227 basis points from a year earlier, against just 6 basis points for conventional loans and 31 for VA. That is a roughly thirty-fold gap between the loan book used by lower-income and first-time buyers and the conventional book. Non-housing consumer debt is deteriorating in parallel across credit cards, student loans and auto loans, and HELOC borrowing is rising again, which is households converting home equity into cash flow. Commercial real estate remains separately weak against a maturity wall of well over $1.5 trillion by the end of 2026. Aggregate housing data therefore continues to look sturdy while the marginal borrower is visibly failing — which is the specific pattern that makes this a slow smash rather than a crash, and also the reason it is easy to miss in national averages.

Sources: HousingWire: Q2 2026 mortgage delinquencies Wolf Street: HELOCs and Q2 delinquencies National Mortgage News: consumer stress CFPB: mortgage performance trends

Status June 3, 2026: worse

The AI/data-center risk has become more physical and financial. Reuters reported on June 3 that private infrastructure and real estate capital are expected to play a larger role in financing the data-center boom, while Goldman sharply raised its hyperscaler capex forecast. Reuters/Devdiscourse: data-center financing

Update June 17, 2026: worse. Compared to June 3: the risk picture is worse. Axios reports Nvidia’s large bond sale and broader AI financing through the debt market. That confirms the point: AI is physical capex, electricity, and borrowing, not just software. Axios: Nvidia bond sale and AI debt

Sources: Axios: Nvidia bond sale and AI debt IMF GFSR April 2026, Box 1.3 Reuters/Devdiscourse: data-center financing

Update June 24, 2026: worse. Compared to June 17: the AI track is more debt-driven. Nvidia bonds and new large compute financings show the boom pulling capital markets deeper into the buildout. That confirms that AI risk here is about power, capex, cash flow, and financing.

Sources: Axios: Nvidia bond sale and AI debt Axios: Anthropic compute financing IMF: GFSR April 2026

Update July 1, 2026: same. Compared to June 24: the AI/data-center risk is unchanged. Slightly lower long rates help financing at the margin, but the core issue remains capex, power, grid capacity, debt, and whether cash flow can carry the buildout if energy or credit conditions tighten again.

Sources: Axios: Nvidia bond sale and AI debt Axios: Anthropic compute financing FRED: 10-year Treasury

Update July 29, 2026: worse. Compared to July 1: the physical scale, power demand, and financing exposure have all increased. Meta expanded its Louisiana data-center plan to 5 GW and more than $50 billion of investment, while Alphabet raised its 2026 capex plan by another $15 billion after strong cloud growth. The IEA now expects global electricity-demand growth to accelerate while Middle East gas disruptions raise generation costs, with data centers among the demand drivers. Bond investors are also beginning to focus more closely on hyperscaler borrowing. These figures do not prove overbuilding, but they make the section's capex-power-debt transmission channel larger and more rate-sensitive than on July 1.

Sources: Reuters: Meta data-center expansion Reuters: Alphabet capex increase IEA: Electricity Mid-Year Update 2026 Axios: hyperscaler debt scrutiny

Update August 19, 2026: worse. Compared to July 29: the build accelerated and its financing became visibly more systemic. J.P. Morgan estimates hyperscaler capex reaches $697 billion in 2026, and the Big Five — Amazon, Microsoft, Google, Meta and Oracle — are spending over $600 billion on infrastructure this year, a 36% increase on 2025, with roughly $450 billion of that, about three quarters, aimed at AI. Demand is genuinely there on the revenue side: second-quarter 2026 cloud growth ran at Google Cloud +82%, Azure +43% and AWS +37%, so this is not a build with no customers. The risk sits in how it is funded. Convertible issuance doubled year over year to $34 billion early in 2026, Morgan Stanley expects $250–300 billion of issuance from hyperscalers alone this year, and projections point to $1.5 trillion of debt over the coming years. The concentration figure matters most: the combined weight of Meta, Alphabet, Amazon and Oracle in the Bloomberg US Corporate investment grade index nearly doubled from 2.2% to 4.1% in the year to April 1, 2026. Ordinary investment grade bond funds therefore now carry materially more AI-cycle exposure than a year ago, whether or not their holders intended it. Meanwhile the binding constraint has moved from chips to power availability, grid capacity and permitting, with over $1 trillion of power infrastructure implied. Hype on top of debt and energy was the framing here; all three legs grew.

Sources: J.P. Morgan: financing AI infrastructure BIS: on- and off-balance sheet borrowing Breckinridge: AI capex and balance sheets Introl: hyperscaler capex 2026

Status June 3, 2026: worse

More indicators are pointing the wrong way at the same time: private-credit losses, CRE delinquencies, yen stress, and renewed oil pressure. That makes the checklist more important, not less. Reuters/Kitco: private credit MBA: CRE delinquencies Reuters/Investing.com: yen

Update June 17, 2026: worse. Compared to June 3: watch the dashboard harder. The most important indicators now are Cushing/inventories, actual Hormuz traffic, insurance, MOU text, high-yield/private credit, yen, and AI debt. One indicator alone is not enough; the danger is several flashing at once.

Sources: EIA: Cushing stocks FRED: high-yield spread FSB: financial-stability vulnerabilities Axios: AI debt

Update June 24, 2026: worse. Compared to June 17: the indicator list flashes harder on energy. Cushing is at 20.03 million barrels, high yield is still calm, and the 10-year is around 4.5%. This is not total panic, but several gauges say the calm is fragile.

Sources: EIA: Cushing stocks FRED: high-yield spread FRED: 10-year Treasury

Update July 1, 2026: same. Compared to June 24: the indicators are more mixed, not solved. The latest official Cushing level is 18.96 million barrels, worse than the June 24 level. At the same time, high yield is 2.80% and the 10-year is 4.38%, which does not show open bond or credit panic. The conclusion is unchanged: physical data and credit data have to be read together.

Sources: EIA: Cushing weekly stocks EIA: Weekly Petroleum Status Report FRED: high-yield spread FRED: 10-year Treasury

Update July 29, 2026: worse. Compared to July 1: more of the dashboard has deteriorated, but it still does not show market-wide panic. Cushing fell from 18.96 to 18.599 million barrels; the 10-year rose from 4.38% to 4.65%; the yen weakened from about 161.83 to 163.69 per dollar; the 30-year mortgage rate reached 6.58%; and Reuters found 28 of 53 listed BDCs loss-making. Counterweights: high-yield was only 2.84%, VIX was 18.21, the St. Louis Fed Financial Stress Index was -0.8263, and the Chicago Fed NFCI was -0.554; negative readings in the last two indicate below-average stress and relatively easy financial conditions. The correct read is worsening physical, rate, and private-credit pressure beneath calm public-market stress gauges.

Sources: EIA: weekly petroleum data FRED: 10-year Treasury FRED: high-yield spread FRED: VIX FRED: financial stress FRED: financial conditions

Update August 19, 2026: same. Compared to July 29: the list is still the right list, so this update simply records where each item now stands. Physical energy: Kpler counted 84 Hormuz transits in a week against more than 100 per day before the war; Cushing rebuilt to 22.6 million barrels; Brent about $92.42. Inflation: US CPI 3.4% with core at 2.5% but energy still 14.7% higher year over year; the euro area back up to 2.9% with energy inflation at 10.0%. Rates: the US 10-year at 4.75%, a 19-month high, the 10y–3m spread near 0.96 points, and Japan's 10-year at 2.95%, the highest since 1996. Credit: high yield at 2.71% and calm, set against BDC unrealised losses of 2.35% of NAV and mortgage delinquencies up 44 basis points year over year. Real economy: payrolls −23,000, retail sales −0.6%, sentiment at 51, gasoline at a record seasonal high. The instructive part is the disagreement between them: the tradable indicators are the calm ones, the physical and household indicators are the stressed ones. Anyone following only credit spreads and the index level would have seen essentially nothing happen this month. That is the whole argument for keeping this list long and mixed.

Sources: EIA: Weekly Petroleum Status Report CNBC: CPI July 2026 FRED: 10-year Treasury FRED: high-yield spread BLS: Employment Situation, July 2026

Status June 3, 2026: worse

Derivative risks have not become smaller. The latest BIS OTC statistics still show enormous notional size, and the FSB warned on June 1 about leveraged strategies in bond markets, exactly the kind of environment where margin calls and collateral scrambles can become dangerous. BIS: OTC derivatives statistics FSB: leveraged trading risks

Update June 17, 2026: same. Compared to June 3: still latent risk. Derivatives become most dangerous when fast price moves, margin calls, and collateral demand arrive at the same time. Energy, rates, and credit remain plausible triggers for that kind of liquidity scramble.

Sources: BIS: OTC derivatives statistics BIS Data Portal: OTC derivatives FSB: leveraged trading risks

Update June 24, 2026: same. Compared to June 17: unchanged latent risk. Nothing in the latest spreads shows derivatives panic, but the FSB warning on leverage and concurrent vulnerabilities keeps margin-call risk relevant if energy, rates, or credit move quickly.

Sources: BIS: OTC derivatives statistics FSB: leveraged trading risks

Update July 1, 2026: same. Compared to June 24: unchanged latent risk. High-yield and rates do not show a margin-call shock today, but Cushing below 20 million barrels and the FSB leverage warning keep the collateral-scramble channel relevant if energy, rates, or credit move quickly.

Sources: BIS: OTC derivatives statistics FSB: leveraged trading risks FRED: high-yield spread

Update July 29, 2026: same. Compared to July 1: derivatives remain a large latent liquidity and counterparty channel, but current market data do not confirm a margin-call shock. VIX was 18.21 and high-yield spreads 2.84% on July 28, while the St. Louis Fed Financial Stress Index was -0.8263 and the Chicago Fed NFCI -0.554 in the week of July 24—both below normal stress levels. Higher long rates and volatile oil preserve plausible triggers, and the FSB's leverage warning remains relevant, but potential contagion should not be described as current contagion without evidence of collateral or funding disruption.

Sources: FRED: VIX FRED: high-yield spread FRED: financial stress FRED: financial conditions FSB: leverage vulnerabilities

Update August 19, 2026: worse. Compared to July 29: no derivatives-specific statistics were published in this window — the BIS semiannual survey is the reference series and has not been updated since the figures cited above — so this update reports the input rather than inventing a headline notional number. What did change is the volatility of the underlying. The 10-year Treasury reached 4.75% on August 18, a 19-month high, the 30-year hit its highest level in nearly two decades, and the 2s/10s curve steepened to about 52 basis points, its widest since May. Interest rates are by far the largest single category of notional exposure in the derivatives complex, so a sustained repricing at the long end changes collateral requirements and hedge values across the system whether or not any individual position fails. The concentration point from the AI sections applies here as well: as Meta, Alphabet, Amazon and Oracle moved from 2.2% to 4.1% of the Bloomberg US Corporate investment grade index, the index and hedging exposure attached to that block grew with it. Nothing has broken, and this update should not be read as claiming otherwise. But the largest table in the casino is interest rates, and that table moved.

Sources: BIS: OTC derivatives statistics Bloomberg: bond rout Breckinridge: IG index concentration FRED: 2s/10s spread

Update June 24, 2026: better. Compared to the June 17 main text: there are better signs in physical traffic and rerouted flows, so the headline picture is not as dark. But the IEA still says a full recovery will not be immediate. This is improvement, not proven molecule peace.

Sources: IEA: supply readjustment after Hormuz shock IEA: Oil Market Report June 2026

Update July 1, 2026: same. Compared to June 24: still improvement in headline tone, not proven molecule peace. The latest official inventory data still leave Cushing at 18.96 million barrels, so the MOU test remains physical flows, insurance, ports, and inventories rather than the existence of a PDF or calmer headlines.

Sources: EIA: Cushing weekly stocks EIA: Weekly Petroleum Status Report IEA: supply readjustment after Hormuz shock

Update July 29, 2026: worse. Compared to July 1: the MOU's central promise of an immediate and permanent end to military operations has failed the real-world test. The United States launched new strikes on July 7 after attacks on tankers; S&P Global counted only 11 Hormuz transits on July 12, versus more than 100 vessels per day before the war; and renewed missile attacks and strikes on July 28–29 again disrupted the pause. Cushing at 18.599 million barrels adds a physical inventory confirmation. Talks and temporary pauses may still reduce risk, but the document has not produced sustained vessel flow or durable peace.

Sources: Reuters: July 7 strikes S&P Global: 11 Hormuz transits AP: renewed attacks July 29 AP: MOU text and resumed fighting

Update August 19, 2026: worse. Compared to July 29: this section's thesis — that a document is not a molecule — received its cleanest demonstration yet. On August 11, US Energy Secretary Chris Wright said the seven-day average leaving Hormuz was almost 9 million barrels per day, plus 5–7 million more via bypass routes. Kpler, working from satellite imagery and vessel transponders, counted 84 transits for the entire week and estimates about 5 million barrels per day by sea; Commodity Context puts it near 7 million by sea and 4 million through pipelines. Kpler's Matt Smith said it is “not possible to reconcile the disparity between what we see and what he is quoting”. CNBC reported traffic near a three-month low on August 12 with the US–Iran deal in doubt, and Iran continues to describe the strait as shut. The legal scaffolding is temporary rather than settled: OFAC's Iran-related General License X runs only to 12:01 a.m. EDT on August 21, 2026, while Treasury has issued its eighth sanctions action of 2026 against Iran's alleged shadow-banking network. Headline peace has now been measured directly against tracked vessels and found wanting, which is precisely the test this section proposed — and the first time the gap has been quantified by named commercial trackers rather than inferred.

Sources: CNN: administration claims vs tracking data Reason: the Hormuz numbers don't add up CNBC: traffic near three-month low OFAC: Iran-related General License

Update June 24, 2026: worse. Compared to the June 17 main text: Cushing fell from 21.64 to 20.03 million barrels for the week ending June 12. That is exactly the red zone this section warns about. If the next weeks do not turn, the inventory argument becomes much sharper.

Sources: EIA: Cushing weekly stocks EIA: Weekly Petroleum Status Report

Update June 25, 2026: worse. Compared to June 24: Cushing fell again, from 20.03 to 18.96 million barrels for the week ending June 19. That puts it below the 20-million-barrel line this section called acute warning territory. EIA’s WPSR also shows commercial crude stocks excluding the SPR down 6.1 million barrels, while the SPR fell further to 331.19 million barrels. Gasoline and distillate stocks rose, which is an offset, but it does not solve the continuing drawdown in crude buffers.

Sources: EIA: Cushing weekly stocks EIA: SPR crude stocks EIA: Weekly Petroleum Status Report

Update July 1, 2026: same. Compared to June 25: no newer official EIA week is in the material before today’s next WPSR release. The latest published Cushing level therefore remains 18.96 million barrels for the week ending June 19, with the EIA page listed as release June 24 and next release July 1. That does not mean conditions are stable; it means the next datapoint is unusually important.

Sources: EIA: Cushing weekly stocks EIA: Weekly Petroleum Status Report

Update July 29, 2026: worse. Compared to July 1: crude buffers have weakened further. Cushing was 18.599 million barrels in the week ending July 24, down 0.361 million barrels, or 1.9%, from the 18.96-million level used on July 1. The SPR fell from 331.19 million barrels in the June 19 report to 307.650 million, while commercial crude fell 7.167 million barrels in the latest week to 404.508 million. There are product offsets: distillate stocks rose 1.061 million barrels during the week and gasoline was nearly flat. But gasoline was still 7.5% below a year earlier and distillates 2.6% lower. The system is still using crude buffers faster than the headline calm implies.

Sources: EIA: Weekly Petroleum Status Report, July 29 EIA: stocks table

Update August 19, 2026: better. Compared to July 29: the inventory picture genuinely improved, and this is one of the few places on this page where that can be said about the direction without hedging. EIA's report for the week ending August 7 shows Cushing at 22.6 million barrels, up 1.6 million on the week and well above the July 24 low of 18.599 million — back above the practical 20-million-barrel line that earlier updates treated as a warning threshold. Commercial crude rose 17.4 million barrels to 424.4 million, an unusually large single-week build, and the SPR stood at 298.7 million barrels. Two qualifications stop this being a clean all-clear. First, level against change: 424.4 million is still about 2% below the five-year average for this time of year, so the build repairs a deficit rather than creating a surplus. Second, the IEA warned in mid-August that reopening Hormuz is becoming more pressing because the world is burning through its stockpiles at a rapid pace — a US rebuild is not a global one, and Cushing is a domestic delivery hub, not a world buffer. With those caveats, the specific indicator this section tracks moved the right way, and the reading that was closest to acute in July is no longer there.

Sources: EIA: Weekly Petroleum Status Report EIA: Cushing weekly stocks Vault Report: US oil inventories August 7 CNN: IEA warning on global stockpiles

Update June 24, 2026: worse. Compared to the June 17 main text: more support for the debt thesis. Axios describes Nvidia’s bond sale and separate compute financing for Anthropic through large credit players. AI is even more clearly a capital-markets issue, not only technology optimism.

Sources: Axios: AI debt boom Axios: Anthropic compute financing

Update July 1, 2026: same. Compared to June 24: same AI-capex risk. The 10-year easing helps financing a little, but the core point is unchanged: the buildout still needs debt, power, cooling, grid capacity, and cash flow that can survive tighter credit or more expensive energy.

Sources: Axios: AI debt boom Axios: Anthropic compute financing FRED: 10-year Treasury

Update July 29, 2026: worse. Compared to July 1: financing dependence is more visible and borrowing costs are higher. S&P Capital IQ data cited by Axios show Alphabet, Amazon, Meta, Microsoft, and Oracle had raised nearly $302 billion through July 22 in a mix of debt and equity—not debt alone—while Goldman Sachs counted $489 billion of AI-related corporate bond and loan supply this year. Alphabet reported negative quarterly free cash flow of $5.9 billion while raising 2026 capex guidance to $195–205 billion, and Nebius arranged a $775 million secured debt facility backed by GPU infrastructure and contracted cash flows. Oracle's five-year CDS reached 212 basis points, although Moody's still described most hyperscaler credit metrics as very strong. The shift is therefore from hypothetical financing risk to materially larger funding and credit-market exposure, not to imminent insolvency.

Sources: Axios/S&P/Goldman: AI financing and CDS Reuters: Alphabet capex and cash flow SEC/Nebius: secured AI-infrastructure debt FRED: 10-year Treasury

Update August 19, 2026: worse. Compared to July 29: every element of the phrase — concrete, power, loans — became heavier. Loans: convertible issuance doubled year over year to $34 billion early in 2026, Morgan Stanley expects $250–300 billion from hyperscalers this year, hyperscalers already raised $108 billion in debt during 2025, and projections point to $1.5 trillion of AI-linked issuance over the coming years. The BIS has published specifically on the on- and off-balance-sheet structures financing this, which is itself a signal about where supervisors are looking. Concrete: Big Five infrastructure spending exceeds $600 billion in 2026, up 36% on 2025, with J.P. Morgan estimating $697 billion across hyperscalers. Power: availability, grid capacity and permitting are now the gating factors rather than chip supply, with implied power infrastructure spending above $1 trillion. The transmission channel is the part that makes this systemic rather than sectoral: the combined weight of Meta, Alphabet, Amazon and Oracle in the Bloomberg US Corporate investment grade index nearly doubled from 2.2% to 4.1% in the year to April 1, 2026. And all of it is being financed with the 10-year at 4.75%, a 19-month high — the debt is being raised into the most expensive long-term funding conditions of the cycle, which is what turns a technology story into a balance-sheet story.

Sources: BIS: financing the AI infrastructure boom J.P. Morgan: financing AI infrastructure Breckinridge: capex and tech balance sheets FRED: 10-year Treasury

Update June 24, 2026: same. Compared to the June 17 main text: same warning light. There is no new data point here that changes the picture by itself, but household stress, a high 10-year yield, and higher daily costs keep subprime auto relevant as an early crack.

Sources: New York Fed: Household Debt and Credit FRED: 10-year Treasury

Update July 1, 2026: same. Compared to June 24: same warning light. The 10-year is lower but still high, and there is no new broad household relief in the data used here. Subprime auto remains an early crack to watch rather than the whole systemic story by itself.

Sources: New York Fed: Household Debt and Credit FRED: 10-year Treasury FRED: high-yield spread

Update July 29, 2026: same. Compared to July 1: no newer New York Fed quarterly report is available, and the latest comparable Q1 data do not show a fresh break. Overall household debt delinquency remained 4.8%, auto-loan transitions into delinquency were mostly unchanged, and auto balances were $1.685 trillion. The Fed's separate BHPH study shows why this remains an early-warning segment: its dealer sample represented only about 2% of all auto loans and 5% of the subprime market, but average subprime BHPH rates were around 25%. Higher gasoline and financing costs add pressure since July 1, but the available loan-performance data still support “elevated and concentrated,” not a new systemic auto-credit break.

Sources: New York Fed: Q1 household debt and credit New York Fed: Q1 report Federal Reserve: BHPH subprime auto

Update August 19, 2026: worse. Compared to July 29: the weak-household thesis is being confirmed in the data, with auto sitting inside a broader pattern of non-housing consumer deterioration. Reporting on second-quarter data describes gradual and persistent deterioration across credit card, student loan and auto loan performance at the same time. The clearest quantified marker of the same divide comes from mortgages, where FHA serious delinquencies rose 227 basis points year over year against just 6 basis points for conventional loans — lower-income borrowers failing roughly thirty times faster than prime ones. The macro conditions behind that worsened this month rather than easing: payrolls fell 23,000 in July with May and June revised down by a combined 103,000; average hourly earnings growth slipped to 3.2%, below the 3.4% CPI rate, so real wages are falling; retail sales dropped 0.6%; and the national average gasoline price reached a record seasonal high near $4.03–4.10 per gallon, roughly 30% above a year earlier. Fuel and food are the least postponable items in a subprime household budget, and they are precisely what rose. The mechanism this section described — the weakest borrowers crack first, well before aggregate data turns — is now observable rather than predicted.

Sources: Wolf Street: household debt and delinquencies Q2 2026 National Mortgage News: elevated consumer stress BLS: Employment Situation, July 2026 AAA: fuel prices August 2026

Update June 24, 2026: same. Compared to the June 17 main text: unchanged slow contagion. The FSB private-credit warning and the Hometap case still support the point: when household equity is packaged as yield, the problem can move between consumer risk and credit risk.

Sources: FSB: private credit vulnerabilities Massachusetts AG: Hometap lawsuit

Update July 1, 2026: same. Compared to June 24: unchanged slow contagion. There is no new datapoint here that changes the conclusion, but private credit, household pressure, and home-equity monetization remain connected risk channels when ordinary credit is expensive.

Sources: FSB: private credit vulnerabilities Massachusetts AG: Hometap lawsuit FRED: 10-year Treasury

Update July 29, 2026: worse. Compared to July 1: legal and household-financing pressure is clearer. Hometap now faces four customer lawsuits filed in 2026 alleging that its HEI contracts are mortgages subject to the Truth in Lending Act and state rules; the latest class action was filed June 23. These are allegations, not court findings, so they do not establish that every HEI is unlawful. But they broaden the legal challenge beyond the Massachusetts enforcement action. At the same time, a 6.58% 30-year mortgage rate keeps ordinary home-equity borrowing expensive, preserving the demand-side pressure that can push households toward alternative contracts. No new authoritative market-size series was found after July 1, so the status change rests on legal exposure and financing conditions, not a claimed surge in HEI volume.

Sources: HousingWire: four 2026 Hometap lawsuits Massachusetts AG: Hometap enforcement action CFPB: home-equity contracts overview Freddie Mac: mortgage rates

Update August 19, 2026: worse. Compared to July 29: the two sides of this section moved toward each other. On the household side, HELOC borrowing is rising again — second-quarter housing coverage was explicitly headlined around the return of home equity lines — meaning households are converting accumulated home equity into current cash flow at the same time as real wages fall and fuel sits at a record seasonal high. On the funding side, private credit's marks deteriorated: aggregate BDC unrealised losses reached 2.35% of net asset value in the first quarter, the deepest quarterly decline since the second quarter of 2022; Blackstone Secured Lending Fund posted its largest quarterly NAV drop in six years, to $25.53 per share, driven by markdowns rather than by a jump in non-performing loans; and the spread advantage over broadly syndicated loans compressed from more than 300 basis points on 2017–2018 vintages to under 100 basis points by the first quarter of 2026. The structural concern this section raised was that private capital is increasingly underwriting household equity at precisely the moment when both the equity and the household come under pressure. Mortgage delinquencies up 44 basis points year over year, with FHA up 227, describe the collateral side of that same trade. Neither leg has broken; both got thinner at once.

Sources: Wolf Street: HELOCs and Q2 2026 housing debt Benzinga: BDC valuation doubts PIMCO: BDC redemptions and NAV pressure HousingWire: Q2 2026 mortgage delinquencies

Update June 17, 2026: same. Compared to the original/latest previous text: still an important food signal. The wheat section remains relevant as part of broader food pressure. This is not an immediate empty-shelves claim; it is the risk of more expensive basic food when energy, fertilizer, transport, and weak harvests interact later in the year.

Sources: USDA WASDE, May 12 2026 USDA NASS: Prospective Plantings USDA Crop Progress

Update June 24, 2026: worse. Compared to June 17: the US side is weaker. USDA’s June WASDE cuts all-wheat production to 1.543 billion bushels and ending stocks to 744 million. Crop Progress also shows only 26% good/excellent for winter wheat on June 21. Global supply is a partial offset, but the US signal worsened.

Sources: USDA WASDE June 2026 USDA Crop Progress, June 22 2026

Update July 1, 2026: same. Compared to June 24: the wheat signal is still weak but not newly worse this week. USDA Crop Progress for June 29 shows winter wheat still at 26% good/excellent, the same headline share as the prior report, while harvest has advanced. The June WASDE cut to production and ending stocks therefore remains the main warning.

Sources: USDA WASDE June 2026 USDA Crop Progress, June 29 2026

Update July 29, 2026: worse. Compared to July 1: USDA cut the US balance sheet again. The July WASDE lowered 2026/27 wheat production from 1.543 to 1.536 billion bushels, the lowest since 1970/71, and reduced projected ending stocks from 744 to 722 million bushels, 22% below last year. This is a smaller buffer and a worse relative forecast. There are meaningful crop counterweights: 81% of winter wheat was harvested by July 26, slightly ahead of the 79% five-year average, and 53% of spring wheat was rated good/excellent versus 49% a year earlier. The careful conclusion remains supply and price pressure, especially in hard red winter wheat—not evidence of empty US store shelves.

Sources: USDA: July 2026 WASDE USDA ERS: wheat market outlook USDA NASS: Crop Progress, July 27

Update August 19, 2026: worse. Compared to July 29: the pressure moved upstream, from this year's crop to the inputs that determine next year's. Sulfur prices are up 261.63% year over year, with Russia's export ban locked in through the end of 2026 and roughly half the world's traded sulfur normally moving through Hormuz, where tanker traffic has collapsed by nearly 90%. A fertilizer chief executive warned in late July that the sulfur shortage could stretch into 2027. Nitrogen and phosphate supply were hit directly by the conflict: Iran halted ammonia production and Qatar suspended urea, ammonia and sulfur output after damage to key facilities, while China extended restrictions on phosphate exports — including DAP, MAP and selected NPK blends — through August 2026, now covering an estimated 50–80% of its export volumes. The World Bank has documented fertilizer prices surging as Hormuz disruptions tighten supply. This reaches grain through cost structure: fertilizer represents 33–44% of corn operating costs and 34–45% for wheat, while USDA's projected 2026 average prices — corn $4.20, soybeans $10.30 and wheat $5.00 per bushel — all sit below estimated break-even. Growers caught between below-break-even output prices and record input costs under-apply fertilizer, and the yield consequence of a year of under-application shows up in the following season rather than this one. The careful conclusion is unchanged from earlier updates: this is a cost-and-margin problem building toward a yield problem, not evidence of empty shelves now.

Sources: World Bank: fertilizer prices and Hormuz American Ag Network: sulfur shortage into 2027 Fertilizer prices weekly update, August 17 Farm Progress: input costs and break-even

Update June 17, 2026: same. Compared to the original/latest previous text: offsets exist, but they are not enough for the main case. The positive scenario needs the MOU to become physical normalization, inventories to stabilize, credit to hold together, and AI productivity to outweigh AI debt. Parts of that can happen, but all of it at once remains low probability.

Sources: IMF: WEO press briefing IEA: Oil Market Report May 2026 Federal Reserve: Financial Stability Report

Update June 24, 2026: same. Compared to June 17: the counterweights are a little clearer but not enough. Hormuz flows and replacement supply seem to be moving the right way, but Cushing, credit, and AI debt mean the positive scenario still requires several simultaneous improvements.

Sources: IEA: supply readjustment EIA: Cushing stocks Axios: AI debt

Update July 1, 2026: same. Compared to June 24: the counterweights still exist but are not enough for a clean base case. Lower long rates and calm public spreads help, while Cushing below 20 million barrels and AI/private-credit financing risks still require several simultaneous improvements.

Sources: FRED: 10-year Treasury FRED: high-yield spread EIA: Cushing weekly stocks

Update July 29, 2026: same. Compared to July 1: the counterweights are better documented, but so are the obstacles to a clean outcome. The IMF projects 3.0% global growth in 2026 and says the world economy has so far weathered the war shock better than feared. US headline CPI fell 0.4% month over month in June, core inflation was 2.6% year over year, and public stress gauges remain calm. Against that, the IMF forecast assumes Hormuz starts reopening in mid-July and normalizes by March 2027; renewed fighting, 11 transits on July 12, Cushing at 18.599 million barrels, and a 4.65% 10-year make that path less secure. The upside case remains real but requires several things to improve together. The original probability weights remain subjective judgments and are not statistically derived, so this update does not pretend to recalibrate them precisely.

Sources: IMF: July 2026 World Economic Outlook Update IMF: assumptions and resilience BLS: June CPI FRED: financial stress EIA: petroleum stocks

Update August 19, 2026: better. Compared to July 29: more items on this list actually went right, and that deserves recording as plainly as the bad news elsewhere. Inventories rebuilt: Cushing rose to 22.6 million barrels for the week ending August 7, back above the practical 20-million line, with commercial crude up 17.4 million barrels to 424.4 million. Inflation eased: US CPI slowed to 3.4% from 3.5%, with core at 2.5% and headline up only 0.1% on the month; energy fell 1.5% month over month. Credit stayed calm: the high-yield spread tightened to 2.71%, investment grade absorbed a $56 billion issuance week without disruption, and agencies expect the US leveraged loan default rate to fall to about 3.0% by October 2026 from 5.3% a year earlier. Employment held up in level terms: unemployment was 4.1% despite the payroll decline. Demand is real in the AI build: second-quarter cloud growth ran at Google Cloud +82%, Azure +43% and AWS +37%, so the capex has customers behind it. Consensus is not forecasting recession: economists put the probability near 25% over the next twelve months, with GDP growth around 2.0%. None of this cancels the energy, household and long-rate deterioration recorded elsewhere in this update. But a risk report that only ever moves one direction is not measuring; this section is the check on that, and this month it moved the other way.

Sources: EIA: Weekly Petroleum Status Report CNBC: CPI July 2026 FRED: high-yield spread Moody's: leveraged finance 2026 U.S. Bank: monthly economic outlook

Update July 1, 2026: same. Compared to the June 24 main text: unchanged system-risk channel. Public high-yield spreads are not in panic, but the FSB and Bank of England framing still matters because private markets can transmit stress through valuations, liquidity gates, refinancing, banks, insurers, and pensions.

Sources: Bank of England: private markets SWES stress scenario FSB: private credit vulnerabilities FRED: high-yield spread

Update July 29, 2026: worse. Compared to July 1: private-market stress has become more visible, while system-wide transmission remains unconfirmed. Reuters' standardized review found 28 of 53 listed BDCs loss-making in the first quarter, versus 12 a year earlier, and average profit across the group moved from $26 million to a $7.6 million loss. A 4.65% 10-year Treasury worsens refinancing and valuation pressure. However, high-yield spreads at 2.84%, STLFSI4 at -0.8263, and NFCI at -0.554 still show calm public credit and below-average broad stress. The Bank of England/FSB transmission channel is therefore more strongly supported at the fund and balance-sheet level, but it has not yet become a confirmed 2007-style system event.

Sources: Reuters: listed BDC losses Bank of England: private-markets SWES FSB: private-credit vulnerabilities FRED: high-yield spread FRED: financial stress

Update August 19, 2026: worse. Compared to July 29: the stress test is being run in real time and the first results have arrived. Aggregate unrealised losses across reviewed BDCs reached 2.35% of net asset value in the first quarter, the deepest quarterly decline since the second quarter of 2022. Average profit fell to negative $7.6 million from positive $26 million a year earlier, with 28 of 53 BDCs loss-making against 12 twelve months before, driven by loan markdowns and rising borrowing costs. In mid-August Blackstone Secured Lending Fund recorded its largest quarterly NAV decline in six years, to $25.53 per share, attributed to portfolio markdowns rather than to a surge in non-performing loans — a distinction that matters, because it means the repricing is happening through valuation rather than through default. The economics that justified the asset class have compressed as well: the spread advantage over broadly syndicated loans fell from more than 300 basis points on 2017–2018 vintages to under 100 basis points by the first quarter of 2026. Analysts remain split, and both readings belong here: some describe limited evidence of meaningful markdown, with loan marks still elevated in absolute terms and relative to the syndicated loan market, while others read the latest results as private credit having avoided the worst case and beginning to stabilise. Both can be true at once — stabilisation at weaker marks, with the question of whether the marks are yet honest still open. That open question is exactly the systemic risk this section is about.

Sources: Analysis: 28 of 53 BDCs loss-making Benzinga: doubts over private credit valuations Private Equity Wire: results point to stabilisation FSB: private credit vulnerabilities

Update August 19, 2026: worse. Compared to the July 29 assessment above: the conditional recovery that section rested on has not arrived, and the price signal has gone further than the earlier text anticipated. Sulfur is up 261.63% year over year. The supply mechanics have hardened rather than eased: Russia's export ban remains locked in through the end of 2026; roughly half the world's traded sulfur normally moves through Hormuz, where tanker traffic has collapsed by nearly 90%; Qatar suspended urea, ammonia and sulfur production after damage to key facilities; and Iran halted ammonia output. A fertilizer chief executive warned in late July that the shortage could stretch into 2027, which is a longer horizon than the gradual late-Q3 to early-Q4 recovery S&P's base case described. The second-order squeeze also materialised: China extended phosphate export restrictions — DAP, MAP and selected NPK blends — through August 2026, now covering an estimated 50–80% of its export volumes, so the same buyers face constrained sulfur and constrained phosphate at once. The World Bank has documented the resulting fertilizer price surge. The earlier caveats still hold and should not be dropped: this remains a cost, margin and availability problem rather than proof of a coming physical famine, and the confirmation signals named above — fresh Gulf loadings rather than stranded cargoes, producer operating rates, named benchmarks, and actual curtailments — are still the right things to watch.

Sources: Acres U.S.A.: sulfur shortage and price American Ag Network: shortage could reach 2027 World Bank: fertilizer prices and Hormuz Fertilizer prices weekly update, August 17