Oil's spike just touched the line. The war is now a rate problem.
The day: US strikes on Iranian targets at noon ET sent WTI +5.2% to $90.22 and the 10-year to 4.79% — it touched 4.80%, the highest since January 2025. S&P 500 closed -0.71% at 7,631.
The call (flagship): Own insurance, not direction — SPY Oct 16 puts ~5% OTM (~$730 strike) through the Sept 16 FOMC, premium ≤1% of portfolio. Target SPY $720–730 or VIX >25; dead on an S&P record close above 7,817.
The disagreement: Consensus trades the war. I trade the transmission — oil above $90 with roughly 2-in-3 priced odds of a rate hike makes this a bond-market event, and bonds are what compress equity multiples.
The level that changes everything: A decisive 10-year close above 4.80%. It clips ~10–12% off AI-equity targets and puts S&P 7,500–7,550 in play.
Published ~6:30 PM ET, after the close — all prices are closing prices.
The most important development of the day was not the strike itself — it was that the strike pushed the 10-year Treasury yield to touch 4.80% for the first time since January 2025. US Central Command announced at noon ET that American forces began striking IRGC targets inside Iran, after tankers were hit in the Strait of Hormuz overnight and a US base in Jordan was attacked over the weekend. WTI settled +5.2% at $90.22. Iran's Tasnim news agency said a "decisive operation" of retaliation had begun, with explosions reported over Aqaba, Jordan.
My stance: short-term bearish tilt, expressed through hedges rather than outright index shorts; longer-term neutral. Confidence: medium-high. Two independent signals agree — the front end now prices a September 16 rate hike at roughly 60–66% across trackers, and the VIX finally woke up, jumping about 10% to ~16.3 for its first gain in five sessions. That is a hedge-into-event tape, not a crash tape: the S&P is still only 2.4% below its record close of 7,817.
A second straight down day, led by the high-multiple end of the market.
| Index | Close | Change |
|---|---|---|
| S&P 500 | 7,631.47 | -0.71% (-55 pts) |
| Dow Jones | 52,766.88 | -0.79% (~-421 pts) |
| Nasdaq Composite | 26,099.77 | -1.03% |
| Russell 2000 | 2,920.13 | -1.23% |
Sectors: Energy was the only clear winner, +1.3% (XLE) on the oil spike, joined by utilities +0.8% and health care +0.7% — the classic defensive rotation. The worst were consumer discretionary -1.7% (XLY) and technology -1.5% (XLK), with industrials -1.4%. Nearly two-thirds of US issues declined, but only 184 of 503 S&P members were down at midday — the damage was concentrated in high-multiple growth, not broad.
VIX, in plain terms: The VIX is the expected 30-day swing in the S&P 500 implied by option prices — Wall Street's fear gauge. It jumped about 10% today to ~16.3 (TheStreet's closing wrap put it at 16.34), its first gain in five days. That is still below its long-run average near 19–20. Translation: after a shooting war, a 4.80% yield touch and a Fed governor openly talking about hiking, options markets are pricing only mild worry. My read (opinion): that is complacency, not calm — and it is exactly why insurance is still cheap.
Treasury yields and what they signal: The 10-year closed at 4.79% (+3–4 bps, highest since Jan 14, 2025, after touching 4.80% intraday). The 30-year finished at 5.27%, near two-decade highs, and the 2-year rose 4+ bps to 4.40% — the front end is now pricing a hike, not cuts. This is global: Japan's 10-year hit 3% for the first time since 1996, German yields are at their highest since 2011, and yields across major markets are at their highest levels since 2008. The signal: heavy government borrowing, an elevated term premium, and — per UBS — rising AI-related debt issuance are competing for capital. Mortgage rates are already at their highest since June 2025.
The ONE key technical level: 10-year yield, 4.80%. It was touched today and closed a hair below (4.79%). A decisive close above it is the regime change: my published rule of thumb is that it compresses AI-linked equity multiples roughly 10–12%, takes ~10–12% off Broadcom-branch targets, and activates the S&P 7,500–7,550 bear band. For the average investor it shows up fastest in mortgage quotes, refi math, and anything financed at the long end.
The catalyst chain ran exactly as the transmission thesis predicted: tanker strikes → CENTCOM strikes at noon ET → WTI +5.2% to $90.22 and Brent +4.6% to $94.65 (both their highest in about six weeks; WTI's biggest one-day gain since July 29) → inflation fear → 10-year to 4.80% → high-multiple equities sold. Roughly a fifth of the world's oil transits Hormuz, and commercial traffic through the strait has already declined sharply in a war that is now six months old. President Trump called the strikes "large and powerful" and threatened a response "at a much harder and higher level" if Iran retaliates again. Iran promptly said it had begun one.
The narrative that got strengthened: September's seasonality now has a mechanism. Monday's review made the case that oil above $90 turns the Fed into the transmission channel. Today stacked the evidence: Fed Governor Michael Barr said that if inflation "appears not to be moderating sufficiently... I think we should act decisively to raise rates"; CME FedWatch puts September 16 hike odds around 60–66%; the ECB is ~99% priced to hike on September 9; and Barclays now expects two more Fed hikes this year. Even gold fell — futures dropped over 1% to ~$4,429 early, with silver -2.5% — because the market is trading inflation as rate hikes, not as a gold-buying event. That is a real-rates regime, and it disciplines every asset in it.
The narrative that got challenged: "the AI trade is immune to rates." The Nasdaq fell more than the Dow; Barclays noted Big Tech (Mag 7 ex-Tesla) has underperformed the S&P 500 on 92% of days this year — its worst run since 2022 — with multiple compression the primary driver, explicitly analogous to the 2022 hiking cycle. But the challenge is more interesting than "tech down": tonight's after-hours tape showed the market is not abandoning AI, it is differentiating. Dell beat and raised (AI server revenue now seen tripling in FY27, $95B backlog) and jumped ~9% after closing down 6.8% in the regular session; GitLab beat and raised and rallied ~20%. Yet MongoDB fell ~12% after hours despite beating and guiding up, and Palo Alto was roughly flat after a beat. Good news no longer buys rallies by default — numbers have to compound to get paid. That is a late-cycle discipline marker, not a bubble pop.
The overlooked factor: the AI boom has started financing itself in the bond market. Two strategist notes today, from two different banks, are two halves of one loop. UBS listed "rising AI-related debt issuance" among the pressures keeping Treasury yields elevated. Barclays wrote that AI-linked financing could "ease capital expenditure concerns" around hyperscaler capex "running ahead of cash flow" and "unlock demand that otherwise could not be funded." Translation: AI capex has outgrown cash flow, and the gap is being filled with debt — debt that competes with Treasuries and mortgages for the same capital, which is part of why the 30-year sits at 5.27% even as the Fed debates hikes. Almost nobody connects the AI trade to the term premium; it is the same trade, on both sides of the balance sheet (see Conclusion).
Real-world implication: The chain from Hormuz to your mailbox is short: oil >$90 → diesel and freight costs (distillate inventories are ~14% below the five-year average) → CPI → the Fed hikes → the 10-year and mortgage rates. Mortgage rates are already at their highest since June 2025. A second-order effect lands on jobs: August ISM manufacturing slipped to 54.6 from 55.6 (below the ~55.2–55.3 consensus) even as its prices index stayed hot on steel. Growth cooling while price inputs stay hot is the exact mix that forces a central bank to choose, and this Fed has told us which way it leans.
Three winners with real news:
Three losers with straightforward reasons:
The most surprising mover: GoPro, +40% to $1.23. The action-camera maker — a ~$1 stock yesterday — agreed to merge with Starman Optical, a private US maker of optical transceivers for AI data centers, in a $285M deal that gives Starman ~90% of the company, pays holders $1.14/share in cash, and leaves them ~10% of an AI-data-center-optics story. It traded as high as $1.64 intraday (+87%) and has roughly doubled in two days, the first on news that YouTuber Markiplier had become its biggest shareholder. Two signals for the broader trend (opinion): first, "AI data center" remains the single most powerful retail-bid phrase in the market — the stock now trades above its cash-out value, meaning buyers are paying for the AI stub itself; second, the marginal bid in small caps is narrative- and influencer-driven, which is exactly the froth profile that gets punished when the VIX regime changes. Honorable mention: Fervo Energy, the geothermal developer, was up double digits premarket on a 396 MW power deal with Google — the AI-power-demand trade is real; the $1 reverse-merger version of it is the froth.
(1) The actionable play for tomorrow — Broadcom, but only after the sentence is spoken (short-term traders). AVGO reports Wednesday after the close at $369.68, pinned on its 200-day. Do not pre-position: my EV math says the print is worth roughly -4% unpositioned, and June's precedent (a beat and a raised quarter that still fell 12% on a held FY27 number) is why. The trade is one sentence — is FY27 AI revenue guidance raised above $100B or not? Branch A (the raise): buy only above $385, target $455, invalidation weekly close below $355, 1–3 months, high conviction. Branch B (the hold): buy the flush $330–348 only if the Q4 AI guide is at least ~$16.5B and margin guide holds ~66%; target $430, invalidation weekly close below $322, speculative, half size. Full branch logic and bear overrides: today's AVGO print preview.
(2) The contrarian move — spend 1% on insurance most investors refuse to own (short-term traders). Buy SPY October 16 puts about 5% out of the money (~$730 strike), covering the Sept 16 FOMC, premium capped at 1% of the portfolio. Target: SPY $720–730 or VIX above 25. Invalidation: an S&P record close above 7,817. Rationale in plain language: you are not betting on a crash — you are refusing to hold a market 2.4% from its record, unhedged, through a war escalation, Friday's jobs report, and a hike priced at 2-in-3. The VIX at ~16.3 is still near 2026's cheapest insurance even after today's +10%. Expect this premium to decay in the base case; that is the price of sleeping. Honest update on my published call: the entry condition was VIX ≤16, and today's close put the window shut — if you are entering now, size down a third rather than chase. Disclosed conflict: this hedge dies in exactly the scenario where AVGO Branch A pays.
(3) The defensive position — short (or simply underweight) small caps via IWM (short-term traders; long-term investors: the trim version). Already live from $293–297, target $282, invalidation a weekly close above $302, 1–3 months, high conviction. IWM closed at $290.57 — 57 cents above $290, the line where floating-rate credit stress starts showing up in small-cap prices. Plain-language rationale: small companies borrow at floating rates, so a hike raises their interest bill within a quarter, while large caps locked in cheap fixed debt years ago — a hike is a direct margin cut to the Russell 2000 and a non-event for most of the S&P. Long-term investors who don't short: trim small-cap exposure and let the defensive pair that worked today (health care +0.7%, utilities +0.8%) do the protecting. Disclosed conflict: the $282 target runs through the $290 line that also threatens my floating-rate credit position (OBDC) — I keep both deliberately, different horizons.
The flagship trade table (hedge):
| Field | Spec |
|---|---|
| Instrument | SPY Oct 16 puts, ~5% OTM (~$730 strike) |
| Entry | Now, premium ≤1% of portfolio; original VIX ≤16 window closed at today's 16.34 close — size down ~1/3 vs the original plan |
| Target | SPY $720–730 or VIX >25 |
| Invalidation | S&P 500 record close above 7,817 |
| Timeframe | Through the Sept 16 FOMC (exit or roll after) |
| Conviction | HIGH on owning insurance (hike odds ~2/3 + VIX off 2026 lows + buyback blackout ~Sept 12 + September seasonality) — NOT a crash call |
| Audience | Short-term traders; long-term investors can express at half size as portfolio insurance |
The most important event: Friday, 8:30 AM ET — the August jobs report. July payrolls printed a shocking -23k, and a pending annual benchmark revision of -79k still hangs over the series. Friday decides the framing: a second negative print with oil above $90 forces the Fed to hike into a weakening labor market — the stagflation corner — while a strong print confirms the hike but removes the recession leg. Both outcomes move markets violently; neither is clearly bullish. Note the calendar quirk: Labor Day (Sept 7) leaves thin liquidity between NFP and the Sept 16 FOMC.
ONE key price level: S&P 500 7,600. Today's close was 7,631. Above 7,600, the summer range holds and dips are buyable for traders with hedges on. A decisive break below it opens clean air down to my published 7,500–7,550 bear band — and would confirm the Dow's warning, having just closed below its 50-day moving average for the first time since April 10 (per MarketWatch's chart note).
Three radar items:
Also on the calendar: Apple's Sept 9 launch event — Ternus' public debut, the expected foldable iPhone, and the first test of whether consumers absorb memory-driven price hikes. The ECB is ~99% priced to hike the same day.
The least-discussed loop in this market: AI capex, Treasury yields, and AI equity multiples are one reflexive trade, and today both ends moved. The same day Barclays diagnosed Big Tech's 92%-of-days underperformance as multiple compression "in a similar setup to 2022," it also described the fix — AI-linked financing that eases concerns about hyperscaler capex "running ahead of cash flow" and unlocks demand "that otherwise could not be funded." UBS, separately, named "rising AI-related debt issuance" as a pressure keeping bonds sold off. Put the two notes together and you have the loop: the AI build-out has outgrown cash flow; the gap is being filled in the credit market; that issuance competes with Treasuries and mortgages; the resulting term premium raises the discount rate applied to AI equities; which compresses AI multiples; which is why Big Tech lags 92% of days. The AI trade is now partially funding itself through the very market that is its biggest valuation headwind. That is why yields rose last week even despite Treasury's expanded buyback program — the pressure is supply, not just Fed expectations.
What to watch, specifically: (1) investment-grade credit spreads — my working trigger is ~120bp (the last S&P quarterly reading I carry is 105.8bp, up from 82.6bp earlier in the cycle): above it, debt-funded AI capex and M&A arbitrage starts closing, and the financing leg of the AI trade breaks before the earnings leg does; (2) Aon's bond pricing this fall — the $18.1bn all-debt USI deal has a ~6.22% break-even blended coupon, and where it actually prints (below ~5.25% and my AON short is wrong) will tell you exactly what the debt market now charges leveraged corporate risk; (3) the AI-linked issuance calendar itself. This continues the AI debt-supercycle thesis we flagged in June — the difference is that the bond market has now started voting on it.
Actionable version: if you run an AI-heavy book, hedge the financing channel, not just the chips — the VIX at ~16 prices almost no fear while credit-spread risk is entirely unpriced in equity portfolios. The cheapest current expression is the October SPY puts above; the cleanest fundamental tell is IG spreads and the fall's AI-linked issuance calendar. When the financing channel tightens, the AI trade's cost of capital rises before its revenues do — position for that sequencing, not for the headlines out of the Gulf.
Index closes per TheStreet's 4:07 PM ET market wrap: S&P 500 7,631.47 (-0.71%), Dow 52,766.88 (-0.79%), Nasdaq 26,099.77 (-1.03%), Russell 2000 2,920.13 (-1.23%), VIX ~16.34 (+~10%, first gain in 5 days; prior close 14.92 per FRED). Cross-check vs ETF closes (Finnhub, pulled 22:00 UTC / 6:00 PM ET): SPY -0.69%, QQQ -1.27%, DIA -0.72%, IWM -1.14% — direction and magnitude consistent (QQQ fell more than the Composite, consistent with mega-cap-led weakness).
Sector ETF closes: XLE +1.27%, XLU +0.78%, XLV +0.66%, XLP +0.32%, XLRE -0.16%, XLC -0.52%, XLF -0.88%, XLB -1.18%, XLI -1.37%, XLK -1.53%, XLY -1.72%.
Yields (CNBC, ~4:40 PM ET): 10Y 4.792% (+3 bps; intraday high 4.80%, highest since Jan 14, 2025), 30Y 5.267% (+1.8 bps, near two-decade highs), 2Y 4.398% (+4 bps). Oil (settles): WTI Oct $90.22 +5.2% (biggest daily gain since July 29 per FactSet), Brent Nov $94.65 +4.6%. Gold futures ~-1.2% early at ~$4,429; silver -2.5% at ~$65.31. ISM Manufacturing (Aug, 10:00 AM ET): 54.6 vs 55.2–55.3 consensus, from 55.6; eighth straight month of expansion; prices still elevated (steel). JOLTS: in line. BTC $77,189 (-2.2% on the day) — the ~$77K wall held its second macro test through the strikes; crypto detail in this morning's crypto daily.
Single-stock closes (Finnhub): AAPL $325.13 +2.61% | MRNA $154.27 +9.93% | MDT $92.04 +1.53% (intraday high $95.41, +5.2%) | AMZN $254.92 -1.87% | CRWD $215.07 -6.90% | AXON $518.30 -8.52% | PANW $362.09 -5.24% | TSLA $356.09 -3.22% | GOOGL $335.02 -1.28% | DELL $425.00 -6.80% (morning high $463) | GTLB $45.09 -3.12% | MDB $434.21 -4.23% | GPRO $1.23 +40.4% (high $1.64; prior close $0.876) | AON $326.30 +1.49% | FRO $44.32 +1.23% | MRVL $210.39 -0.60% | NVDA $217.44 -1.51% | AVGO $369.68 -0.18%. After-hours (CNBC, as of ~4:52 PM ET): DELL +~9%, GTLB +~20%, MDB -~12%, PANW ~flat, CRDO -~4%.
The framework is the transmission chain published in Monday's review: Hormuz escalation → oil >$90 → diesel/CPI → hike odds → 10Y → equity multiples. Scenarios into Sept 16 (weights carried from Monday, updated): Bull 20% — de-escalation or a deal, Brent back to mid-$80s, 10Y <4.70%, S&P retests 7,816.70 (close IWM short above $302 weekly; puts die). Base 45% — chop, S&P 7,600–7,775, oil $88–95, 10Y 4.70–4.80%, NFP ambiguous; time decay eats the puts, IWM grinds toward $282. Bear 35% — weak NFP plus oil >$95, decisive 10Y close >4.80%, escalation against US regional assets; S&P 7,500–7,550, VIX >25, puts pay, IWM target hit. Load-bearing input: the hike actually happens (or stays fully priced) into Sept 16. If Friday's report is strong-but-not-hot while oil retreats below the mid-$80s, hike odds fall below 40% and the whole hawkish structure unwinds — the answer flips to "close shorts, keep only the AVGO event branches," and I will publish that reversal. Secondary assumption: Hormuz stays disrupted; a signed reopening arrangement is the single fastest falsifier of the oil leg.
EQ-IWM-1 short from $293–297: closed $290.57, target $282, 57¢ above the $290 warning line — mid-move, working. EQ-VIXHEDGE-1 SPY puts: VIX ≤16 entry window closed (VIX 16.34); if entering now, down-size ~1/3. EQ-AVGO-1: branches A/B armed for tomorrow's print (linked piece). EQ-FRO-1 long from $43.80: closed $44.32 (+1.2%); escalation leg live, kill switch = signed Hormuz deal. EQ-AON-1 short-avoid: $326.30 (+1.5%), entry zone $332–348 never filled, no invalidation approached. EQ-MRVL-1 long from $216.62: $210.39, inside the $205–222 zone; Oct 6 Investor Day is the catalyst. EQ-NVDA-2 long from $220.78 mark: $217.44, inside the $205–222 zone. PC-HIKE-1 (OBDC): carried, unchanged. CR-BTC-1: $77,189, wall held its second macro test. CR-ETH-1: opened this morning, see the crypto daily. Disclosed conflicts, unchanged from Monday: the puts pay in the scenario that kills AVGO Branch A and both crypto legs; the IWM $282 target runs through the $290 line that threatens OBDC; four positions lean hawkish with NVDA, the puts and the AVGO branches as deliberate counterweights.