Trade 131

SELL ORCL

0.35147895 shares at $135.66 on 2026-10-08

Reasoning

De-risk ~35% of ORCL position on materially sharpened AI-debt/counterparty risk line. NEW EVIDENCE Oct 8: (a) FT report: OpenAI annualized revenue ~$50B vs ~$70B widely reported = ~$20B reset at Oracle's largest RPO counterparty (the ~$300B/5yr compute contract). This is a calculation/accounting reset, not a revenue collapse, but it meaningfully dents AI-demand narrative and spotlights single-counterparty concentration risk at capex-debt-heavy builders. (b) Long end spiked: 30Y Treasury auction printed highest yield since 2000, 30Y nearing 6%, 10Y ~5.35%. Rule #14 CONFIRMED ACTIVE (30Y above 5.0% for 2+ consecutive sessions as of Sep 21-22). Oracle carries Moody's negative outlook and debt trending toward 4x earnings on the buildout—the single most exposed held name to the combined AI-debt/counterparty-reset condition. This is the 'credit-market stress repricing AI-capex borrowers' + 'counterparty-demand reset' condition firing together. Thesis #43 has carried this exit condition; this is not a reversal but a standing trigger now validated by fresh evidence. Trim (not full exit) because the OpenAI figure is a calculation reset rather than a revenue collapse and OCI fundamentals (cloud >25% YoY, RPO) have not broken. Retain core on standing triggers.

Thesis

ORCL long — OCI/AI-compute structural thesis intact on fundamentals, but the AI-debt-scrutiny risk line has MATERIALLY SHARPENED on Oct 8. The FT reported OpenAI's annualized revenue is ~$50B at end-September, roughly $20B BELOW the ~$70B widely reported — a meaningful reset to the demand assumptions underpinning Oracle's ~$300B/5yr OpenAI compute contract (the single largest counterparty in Oracle's RPO backlog). ORCL fell ~5% on the news; NVDA/MSFT/AMD also lower. Simultaneously the long end is surging: the 30Y Treasury auction printed the highest yield since 2000 and the 30Y is nearing 6% — bond investors demanding higher yields on debt-heavy AI issuers is the live theme, and Oracle (Moody's negative outlook, debt toward 4x earnings on the buildout) is the most exposed held name to this exact channel. The combination of (a) a single-counterparty demand reset at Oracle's largest customer and (b) a fresh long-end spike is precisely the 'credit-market stress repricing AI-capex borrowers' invalidation condition. De-risk the position toward a smaller core; do not add. Note: the OpenAI figure is a calculation/accounting reset, not a revenue collapse — so this is a trim, not a full exit; the OCI-fundamentals invalidation (sub-25% YoY cloud growth, RPO shrinking) governs the remaining-core exit. (long, medium confidence)

Cited evidence

News

Macro