1HeadlinesWhat does this mean?
Stage 3 (forced deleveraging), up from Stage 2 in the last report; earliest downgrade 23 November. AI economics: Elevated, unchanged.
The bad
- Confirmed hedge-fund deleveraging. Goldman Sachs prime-brokerage data showed all three major hedge-fund strategies it tracks lost more than 1% on 22 October, the first simultaneous decline of that magnitude since the COVID-era market turmoil.1
- Through 22 October, technology-focused hedge funds had lost approximately 10% during October and Asia-focused funds approximately 9%.1 Crucially, the selloff prompted hedge funds to reduce leverage and risk exposure.
- New central-bank evidence of potential international credit contagion. The Bank of Japan’s Financial System Report explicitly connected increasing AI infrastructure investment financed through private funds and other NBFIs, the growing exposure of Japanese banks to foreign non-bank lenders and investment funds, and the possibility that declining AI profitability triggers portfolio adjustments that amplify market losses internationally.2
The good
- Japan’s banking system was still judged sufficiently capitalised in the Bank of Japan’s stress tests.2
- Ordinary credit has not followed: the high-yield spread was 338bp and the investment-grade spread 91bp at the 23 October close, both inside Calm.312
- It is not yet proven that the deleveraging will generate forced asset sales, funding withdrawals or material losses for banks and private-credit lenders.
2ScorecardWhat does this mean?
| Pillar | Status | Trend | Why |
|---|---|---|---|
| AI equities | Elevated | Unchanged | The semiconductor index closed 26.3% below its 52-week high on 23 October, inside the Elevated band and beyond the Stage 1 tripwire.4 The report’s equity evidence is the deepening tech selloff behind October’s hedge-fund losses: technology-focused funds down about 10% and Asia-focused funds about 9% for the month through 22 October.1 |
| Rates and macro | Elevated | Unchanged | US 10-year at 5.00%, inside the Elevated band.7 No new rates evidence in this report; status held from the 24 October report. |
| Broad credit (the firewall) | Calm | Unchanged | High-yield spread 338bp and investment-grade 91bp, both inside Calm, though the high-yield spread is now 12bp from the Watch threshold.312 The potential international contagion the report flags comes from a central-bank warning, not from wider spreads. |
| AI-specific credit | Stressed | Unchanged | No new AI-specific credit evidence in this report; status held from the 24 October report. |
| AI financing flow | Elevated | Unchanged | No new AI financing evidence in this report; status held from the 24 October report. |
| Leverage and liquidity | Stressed | Worse | Goldman Sachs prime-brokerage data showed all three major strategies it tracks losing more than 1% on 22 October, the first simultaneous decline of that size since the COVID-era turmoil, and the selloff prompted funds to reduce leverage and risk exposure. That is evidence of actual deleveraging across strategies, not simply accumulating portfolio losses.1 |
| Private credit | Elevated | Unchanged | No new private-credit market evidence in this report; status held from the 24 October report. The Bank of Japan’s warning about AI infrastructure financed through private funds and other NBFIs, and Japanese banks’ exposure to foreign non-bank lenders, is supervisory context rather than a scored event.2 |
| AI economics | Elevated | Unchanged | No new capex or cash-flow evidence in this report; status held from the 24 October report. The Bank of Japan names declining AI profitability as the trigger for the portfolio adjustments it fears, but that is a scenario, not a reading.2 |
3Change since last reportWhat does this mean?
Compared against the 24 October 2026 report, back-scored the same way. Scores for this report were back-scored on 4 November 2026 from its own text and from FRED and Yahoo Finance closing data, because it predates the scorecard format.
| Last report (24 Oct) | This report | Change | |
|---|---|---|---|
| Stage | 2 (evidence 23 Oct) | 3 (evidence 23 Oct) | Up. Goldman prime-brokerage data showing synchronised hedge-fund losses on 22 October, and funds deliberately cutting leverage and risk exposure (reported 23 October), meets the Stage 3 entry condition.1 |
| AI economics | Elevated | Elevated | Unchanged. No new economics evidence. |
| AI equities | Elevated | Elevated | Unchanged. Same 23 October close as the last report. |
| Rates and macro | Elevated | Elevated | Unchanged. Same 23 October close as the last report. |
| Broad credit | Calm | Calm | Unchanged. Same 23 October close as the last report; spreads still inside Calm.3 |
| AI-specific credit | Stressed | Stressed | Unchanged. No new evidence. |
| AI financing flow | Elevated | Elevated | Unchanged. No new evidence. |
| Leverage and liquidity | Elevated | Stressed | Worse. Portfolio losses have turned into documented deleveraging across strategies.1 |
| Private credit | Elevated | Elevated | Unchanged. No new evidence. The Bank of Japan report adds supervisory context on non-bank AI financing.2 |
Net: one pillar moved, and it moved the stage. Leverage and liquidity went from Elevated to Stressed on the Goldman prime-brokerage data, which takes the headline from Stage 2 to Stage 3. Nothing else changed status: the market data is the same 23 October close the last report used, and broad credit is still Calm.
4Transmission chainWhat does this mean?
- 1 AI equity correction Confirmed
- 2 AI-specific credit stress Confirmed
- 3 Forced deleveraging Confirmed 23 Oct Earliest downgrade 23 Nov
- 4 Firewall breached Partial evidence
- 5 Systemic liquidity event Not present
| Stage | Status |
|---|---|
| 1 — AI equity correction | Confirmed. The semiconductor index has been more than 20% below its 52-week high for five consecutive closes since 20 October, and closed 26.3% below it on 23 October.4 |
| 2 — AI-specific credit stress | Confirmed. In place since 22 September and last refreshed by the 23 October evidence in the 24 October report. No new AI credit evidence in this report. |
| 3 — Forced deleveraging | Confirmed. Entry evidence 22–23 October: Goldman Sachs prime-brokerage data showing all three major strategies down more than 1% on 22 October and funds cutting leverage and risk exposure, reported 23 October.1 Earliest possible downgrade is 23 November. Not yet shown: whether the deleveraging produces forced asset sales, funding withdrawals or material losses for banks and private-credit lenders. |
| 4 — Firewall breached | Partial. High-yield and CCC spreads have been widening (338bp and 1,356bp) but remain far below the tripwires; investment-grade is 91bp against 130bp.31312 |
| 5 — Systemic liquidity event | Not present. The Bank of Japan sets out the international route, AI-driven portfolio adjustments amplifying losses through banks’ exposure to foreign non-bank lenders, but still judged Japan’s banks sufficiently capitalised.2 |
5JudgementWhat does this mean?
The Goldman Sachs prime-brokerage data changes the previous assessment. All three major strategies losing more than 1% on the same day, for the first time since the COVID-era turmoil, would on its own be another report of accumulating losses. What makes it different is that the selloff prompted hedge funds to reduce leverage and risk exposure: there is now evidence of actual deleveraging across investment strategies, rather than simply portfolio losses piling up.1
The Bank of Japan’s Financial System Report supplies the map for where that could go. It explicitly connects AI infrastructure investment financed through private funds and other NBFIs, Japanese banks’ growing exposure to foreign non-bank lenders and investment funds, and the possibility that declining AI profitability triggers portfolio adjustments that amplify market losses internationally. It still judged Japan’s banking system sufficiently capitalised in stress tests.2 The most significant development is that the portfolio-adjustment mechanism identified in central-bank warnings was already operating among hedge funds.
What remains unproven is whether that deleveraging generates forced asset sales, funding withdrawals or material losses for banks and private-credit lenders. Until it does, the firewall is intact: broad high-yield and investment-grade spreads are still inside Calm.3
6Key indicatorsWhat does this mean?
Market data as at the close on 23 October 2026, compared with the close used by the 24 October 2026 report (23 October 2026). Reported figures are carried forward from the report that last cited them until a newer one appears.
| Indicator | Latest | Previous | Change | Tripwire | Source |
|---|---|---|---|---|---|
| AI equities | |||||
| PHLX Semiconductor Index (SOX) | 10,779; 26.3% below its 52-week high (22 Jun) | 10,779 | ±0.0% | Stage 1 when more than 20% below its 52-week high for five closes | Confirmed, Yahoo Finance ^SOX4 |
| Nasdaq 100 | 28,958 | 28,958 | ±0.0% | Context | Confirmed, Yahoo Finance ^NDX5 |
| KOSPI | 5,550 | 5,550 | ±0.0% | Context | Confirmed, Yahoo Finance ^KS116 |
| Rates and macro | |||||
| US 10-year Treasury | 5.00% | 5.00% | ±0.00pp | Elevated 4.75–5.25%; Stressed above 5.25%, each for five closes | Confirmed, FRED DGS107 |
| US 30-year Treasury | 5.33% | 5.33% | ±0.00pp | Context | Confirmed, FRED DGS308 |
| Brent crude | $128.30 | $128.30 | ±0.00 | Context | Confirmed, FRED DCOILBRENTEU9 |
| US dollar index (broad) | 121.32 | 121.32 | ±0.0% | Context | Confirmed, FRED DTWEXBGS10 |
| Yen per US dollar | ¥156.87 | ¥156.87 | ±0.00 | Context | Confirmed, FRED DEXJPUS11 |
| Broad credit | |||||
| ICE BofA US investment-grade spread | 91bp | 91bp | ±0bp | Watch at 100bp; Stage 4 above 130bp for five days | Confirmed, FRED BAMLC0A0CM12 |
| ICE BofA US high-yield spread | 338bp | 338bp | ±0bp | Watch at 350bp; Stage 4 above 450bp for five days | Confirmed, FRED BAMLH0A0HYM23 |
| ICE BofA CCC & lower spread | 1,356bp | 1,356bp | ±0bp | Early warning only | Confirmed, FRED BAMLH0A3HYC13 |
| ICE BofA CCC & lower effective yield | 18.17% | 18.17% | ±0.00pp | Early warning only | Confirmed, FRED BAMLH0A3HYCEY14 |
| AI-specific credit | |||||
| AI investment-grade premium over the broad index | Oracle 2031 ~263bp, SpaceX ~228bp vs BB average ~206bp (24 Oct report) | Oracle 2031 ~263bp, SpaceX ~228bp vs BB average ~206bp | no new reading | n/a | Carried from the 24 October report |
| Oracle 5-year CDS | ~288bp (15 Oct report) | ~288bp | no new reading | n/a | Carried from the 15 October report |
| Meta 5-year CDS | 114.79bp (record) (23 Oct report) | 114.79bp (record) | no new reading | n/a | Carried from the 23 October report |
| CoreWeave 5-year CDS | ~1,200bp (17 Oct report) | ~1,200bp | no new reading | n/a | Carried from the 17 October report |
| SoftBank 5-year CDS | ~424bp (15 Oct report) | ~424bp | no new reading | n/a | Carried from the 15 October report |
| Named data-centre project debt | Hyperion, Sopaipilla and Project Excalibur bonds substantially below par (16 Oct report) | Hyperion, Sopaipilla and Project Excalibur bonds substantially below par | no new reading | n/a | Carried from the 16 October report |
| Quarterly financials | |||||
| Meta 2026 capex guidance | lower end of range raised (22 Oct report) | lower end of range raised | no new reading | n/a | Carried from the 22 October report |
| Meta quarterly free cash flow | negative (Q3) (22 Oct report) | negative (Q3) | no new reading | n/a | Carried from the 22 October report |
Yahoo Finance ^SOX, daily closes to 23 Oct.
FRED DGS10, daily closes to 23 Oct.
FRED BAMLC0A0CM, daily closes to 23 Oct.
FRED BAMLH0A0HYM2, daily closes to 23 Oct.
Faint vertical lines mark each of the last 20 reports. Dashed red lines are the published thresholds.
What would change our view
Would make it worse (up toward Stage 4): Evidence that the hedge-fund deleveraging is generating forced asset sales, funding withdrawals or material losses for banks and private-credit lenders; the high-yield spread above 450bp, or investment-grade above 130bp, for five consecutive trading days.
Would make it better (down to Stage 2): A month with no new forced-selling evidence takes the stage from 3 to 2. The clock started on 23 October, so the earliest that can happen is 23 November, and it restarts if new margin calls, liquidations or prime-broker withdrawals are reported.
Would move the AI-economics axis: Evidence in hyperscaler cash flow, capex guidance or cloud revenue of the declining AI profitability the Bank of Japan names as its trigger.
7SourcesWhat does this mean?
- Reuters — Hedge fund returns worsen this week as tech selloff deepens (2026-10-23) —
reconstructed-match - Bank of Japan — Financial System Report (October 2026) (2026-10-23) —
official-source - Federal Reserve Bank of St. Louis (FRED) — ICE BofA US High Yield Index Option-Adjusted Spread (continuous) —
official-data-series - Yahoo Finance — PHLX Semiconductor Index (^SOX) price history (continuous) —
market-data-series - Yahoo Finance — Nasdaq 100 (^NDX) price history (continuous) —
market-data-series - Yahoo Finance — KOSPI (^KS11) price history (continuous) —
market-data-series - Federal Reserve Bank of St. Louis (FRED) — 10-Year Treasury Constant Maturity Rate (continuous) —
official-data-series - Federal Reserve Bank of St. Louis (FRED) — 30-Year Treasury Constant Maturity Rate (continuous) —
official-data-series - Federal Reserve Bank of St. Louis (FRED) — Crude Oil Prices: Brent - Europe (continuous) —
official-data-series - Federal Reserve Bank of St. Louis (FRED) — Nominal Broad U.S. Dollar Index (continuous) —
official-data-series - Federal Reserve Bank of St. Louis (FRED) — Japanese Yen to U.S. Dollar Spot Exchange Rate (continuous) —
official-data-series - Federal Reserve Bank of St. Louis (FRED) — ICE BofA US Corporate Index Option-Adjusted Spread (continuous) —
official-data-series - Federal Reserve Bank of St. Louis (FRED) — ICE BofA CCC & Lower US High Yield Index Option-Adjusted Spread (continuous) —
official-data-series - Federal Reserve Bank of St. Louis (FRED) — ICE BofA CCC & Lower US High Yield Index Effective Yield (continuous) —
official-data-series
Not investment advice. Scoring, tripwires and the persistence rule are set out in How to read the AI Market Watch; terms are explained in the explainers.