On plan, working

S&P cuts Oracle to BBB-, one notch above junk, naming OpenAI as a key credit risk - the cleanest identifiable forced-seller trigger on the board

What changed: S&P Global lowered Oracle to BBB- from BBB (stable) on July 9, one notch above speculative grade, explicitly naming OpenAI - which accounts for roughly HALF of Oracle's record $638B backlog - as a key credit risk, since OpenAI is private, has never turned a profit, and burns cash. Five-year Oracle CDS hit ~203bp, the highest since the series began in late 2008. The financial picture behind it: FY2026 capex $55.66B against FY2026 free cash flow of about -$23.69B; S&P models FY2027 capex at $90-95B (up from a $60B prior forecast) and a free operating cash-flow deficit near -$42B; total debt ~$167B; adjusted leverage heading to mid-4x against the ~4x S&P treats as BBB-appropriate; funding via ~$20B of equity issuance in calendar 2026 on top of a $5B mandatory convertible, with roughly $40B more debt-and-equity signaled. The stock sits near $122.69, about -65% from its $345.72 peak. Why it matters: this is the single cleanest, most identifiable forced-seller trigger in the market. One more notch takes Oracle to junk, and index-mandated investment-grade holders would then be MECHANICALLY required to sell - a rules-based, non-discretionary flow, which is exactly the kind of catalyst that turns a quiet credit index into a repricing one. Transmission: it is the corporate-credit link in the AI chain the book tracks - Korea to memory to semis to hyperscaler capex to the AI-capex funding gap (Alphabet's equity raise, June 4; the July 22 negative-FCF prints) to AI-linked credit - and it is the concrete mechanism by which the July 23 'index credit is silent' gap could close. Persistence: until the leverage path or the OpenAI concentration changes; the FY2027 capex step-up is guided, not optional. Invalidation: a stabilizing or upgraded outlook, OpenAI concentration falling materially, or the capex plan being cut. Reaches the book indirectly - Oracle is not held - through the credit-cycle-sensitive financials PLMR and PYPL, and it is the credit read on the same AI capex that drives TSM's order book.

Raise stop

Raised the PLMR stop from 101.33 to 114.49 (breakeven) on the remaining 145.98 units, so a winner that ran +21% can no longer round-trip below the entry. Locks the risk to zero on the held lot.

Engine: Raise stop→Executed: Followed

Trimmed 40.02 @ 139.55

Trimmed 40.02 of 186 PLMR units at 139.55 (+21.9% on the lot, +$1,002.90 realized) on the engine's over-extension and single-name concentration call. Reduced PLMR from the book's largest position toward target weight while keeping the core.

Engine: Trim→Executed: Followed

HY OAS at 267bp masks a CCC−B dispersion blowout past 600bp: the public-credit crack the feed had no visibility into

What changed: index high-yield option-adjusted spread sits at 267bp (ICE BofA / FRED, July 7) - late-cycle complacency on the standing framework (below 350bp is complacency, 600bp is stress, above 800bp has coincided with or preceded every US recession since the 1990s). Beneath the calm index, the CCC-and-below minus single-B spread has blown past 600bp, roughly 200bp wider year-to-date (State Street Q3 outlook). Why it matters: dispersion widening while the index stays tight is the textbook credit-cycle inflection - the weakest credits reprice first and the index follows. HY OAS leads equity drawdowns by 2 to 4 weeks in stress regimes because credit dealers reprice ahead of equity vol-targeting flows, so the CCC−B blowout means that lead may already have started. This is the one genuinely NEW, independent signal in the July 13 feed - the book previously carried no public-credit-spread visibility at all. Transmission: it is the listed-market complement to the private-credit stress already logged (the June 20 redemption gates and the July 6 15.6B-request / 38%-fill escalation) - the same credit-cycle turn now visible in public spreads. Private-credit 'true' default (including selective defaults and liability-management exercises) approaches ~5% versus the sub-2% headline, PIK-toggle use is rising, and software is 20 to 30% of private-credit assets against ~5% of public HY - the leveraged expression of the same AI-disruption bet the token-rollover (July 4) and data-center-lease (July 5) events track. Energy is ~14% of HY, and the inverse-tail link to Hormuz matters: a 79-to-100-dollar Brent supply shock compresses energy-HY spreads while crushing every oil-consuming HY issuer (transport, chemicals, airlines) - a stagflationary credit shock, not an energy-credit rally. Persistence: structural through the credit cycle; the primary trigger is HY OAS breaking 350bp, which buys a 2-to-4-week lead to de-risk equity beta, watched alongside the CCC−B differential. Invalidation: dispersion compressing back toward the index, or HY OAS holding under 300bp with the CCC−B gap narrowing. Reaches the book through its credit-cycle-sensitive financials - PLMR (specialty insurer) and PYPL (payments and credit) - more than the broad index.

S&P −1.21% with HY OAS at 268bp: credit is not confirming the equity selloff - the sharpest live test of the credit-leads framework

What changed: the S&P 500 fell 1.21% to 7,408.30 - its largest one-day drop in about a month - while index HY OAS sat at 268bp (July 22 observation, ICE BofA / FRED), essentially unchanged at the tight end of the post-2009 range. Credit is NOT confirming the equity selloff. Why it matters: the standing credit framework (logged July 13) holds that HY OAS leads equity drawdowns by 2 to 4 weeks in stress regimes, with 350bp as the de-risk trigger. Either equities are overshooting a rates-and-multiple story with no solvency component (drawdowns beyond ~5% almost always require credit participation), or credit's lead simply has not started yet and the CCC−B dispersion (already blown past 600bp, July 13) is the early wire. We don't know which yet - that unresolved question IS the event; it is logged as the sharpest live test of the credit-leads framework, not as a verdict. Transmission: if OAS holds sub-290 while equities keep falling, the leading-indicator thesis needs recalibration; if OAS breaks 350, the July 13 playbook activates with its 2-to-4-week de-risk window. Persistence: resolves within weeks, one way or the other. Invalidation: of the equity-overshoot read, OAS through 350bp; of the credit-leads read, equities reclaiming the selloff with OAS never confirming. Reaches the book through PLMR and PYPL, the credit-cycle-sensitive financials.

28 of 53 public BDCs now loss-making and Apollo gates at the 5% cap against 16.8% requested: the private-credit factor is confirmed, not suspected

What changed: the private-credit stress the book has logged twice (June 20 redemption gates; July 6 15.6B-request / 38%-fill) escalates to confirmed with hard Q1 2026 figures - 28 of 53 publicly traded BDCs posted losses versus 12 a year earlier, average profit swinging from +$26M to -$7.6M; the Proskauer Private Credit Default Index rose to 2.73% from 1.84% two quarters earlier (697 loans, $189.2B); Apollo Debt Solutions BDC received repurchase requests for ~16.8% of shares outstanding and honored only the 5% quarterly cap, paying ~$0.7B - the canonical maturity-mismatch datapoint of this cycle; Palmer Square Capital BDC took a $48.3M realized-plus-unrealized loss with NAV/share falling 10.4% in one quarter; and dividends were cut across the complex (TCPC -32%, OBDC -16%, CGBD -13%, FSK -7%). Why it matters: a gate is not a drawdown, it is a liquidity mismatch made explicit - investors want out faster than the vehicles can sell. Transmission: US banks have extended roughly $300B of credit to private-credit funds, BDCs and CLOs, so the channel to the regulated system exists (though at ~1.9% of Tier 1 for the Y-14 sample, it is a stress vector, not yet a solvency one); it is the same credit-cycle turn as the July 13 public-spread dispersion, and it is the leveraged expression of the AI-disruption bet (software is 20-30% of private-credit assets). Persistence: through the credit cycle; watch fill rates, the default index, and further gates. Invalidation: fill rates normalizing toward par and the default index receding. Reaches the book through PLMR and PYPL, the credit-cycle-sensitive financials.

Held: 145.98 @ 133.79

Thesis breaks if: A bad catastrophe quarter.