The shock absorber just got eight percent deep — and nobody is long the tape it protects.

Record high, empty hands

1. The hike is being priced out, one meeting at a time.September now prices barely a one-in-three chance of a hike, and no chance at all of a cut. A month ago the market gave three-in-four odds to at least one hike, including a one-in-four chance of two. Three weeks of disinflation have taken three-quarters of the tightening risk out of the front end, and the index made records the whole way. The full ladder is in Section 09.
2. The gamma flip collapsed in one session.Thursday's close moved the flip from 7,579.15 to 7,136.88 in a single day. The put wall rose a hundred points to 7,600; the call wall held 8,000. That is the mechanical reason a cool PPI and a record close produced a 0.65% day, not a 2% one.
3. Positioning is short into the record.Large speculators are net short 27,258 S&P contracts and 14,639 Nasdaq contracts, having added 20,536 Nasdaq shorts against 983 longs in the latest reporting week. Retail agrees: the AAII spread is −3.2. A record the futures market is short and the retail survey disbelieves is a squeeze setup, not a distribution.
4. Beats are still not being paid for.Applied Materials beat both lines and guided the October quarter to $10.25bn against $9.55bn consensus. It closed down 2.48% and lost another 5.21% overnight — the fourth AI-adjacent name in a week to print well and be sold.
5. The cool PPI had a hot core inside it.Headline producer prices were flat against +0.2% expected and decelerated to 4.7% annually. The Bureau's narrow core — excluding food, energy and trade services — ran +0.4% against +0.2% expected. The market traded the headline. Nobody traded that line.

Where everything sits

Single names — Thursday and the overnight

Index and single-name closes are Thursday's cash session; futures, commodities, FX and overnight quotes are live at 11:15–11:30 PM ET Thursday. Settlements are Cannon's CQG marks.

Sentiment & flow gauges

The flow read

Four instruments, one message, and it is not the message a record close implies. Speculators are net short both equity contracts; retail is net bearish; and per work circulating from Bank of America's systematic desk, index hedgers carry roughly 22,000 short contracts concentrated in the 7,750–7,900 strike band — precisely where cash is trading — with trend followers rebuilding length and no sell trigger short of a four-percent drawdown. This market did not buy the high. It hedged it, and is watching the hedge cost money.

Yesterday's calls, graded

One consumer print, then the weekend

Neither branch is a range problem — with the flip 662 points below spot, hedging damps the day regardless. The branch decides the direction of a small day, and which end of the curve leads into next week's minutes.

Pivots, walls and the shape of the curve

Cannon Daily Levels — August 14, 2026 · September contracts

Cannon Edge — daily futures snapshot · closes are Cannon's own CQG settles

Dealer gamma map — SPX and ES

Two things changed in one session. The flip fell 442 points, putting a public dealer-gamma model deep in positive gamma — hedging sells strength and buys weakness, and the market would have to fall more than eight percent before that reverses. Meanwhile the put wall moved up a hundred points, tightening the floor. The walls now sit almost symmetrically around cash, roughly two and a half percent either side — a narrower box than the market has traded in for weeks. The ES premium re-derives daily; translate with today's 23.51.

Cannon pivots — ES September

Pivot 7,808.42, resistance 7,852.58 / 7,882.67 / 7,926.83, support 7,778.33 / 7,734.17 / 7,704.08. Futures sit between the pivot and R1, and R1 lies just above the contract high of 7,838.50 — so the first real resistance is the record itself, sixteen points overhead, and the pivot is where buyers must show up to keep the structure intact.

The VX curve

August 16.00, September 18.11, October 19.60, against Thursday's spot close. Spot-to-front +1.37, front-to-second +2.11 — steep contango, no backwardation. September straddles the FOMC and carries a two-point premium to August for it: a market pricing the meeting as an event, not a threat.

New and moved voices

The full roster

Ordered by weight on today's setup. Unchanged views appear here only; new and moved voices get the full treatment in Section 05.

What the data actually said

The producer print was two numbers pointing opposite ways and the market traded one. Headline final demand was flat against +0.2% expected, the annual rate down to 4.7% from 5.5%. Underneath, the Bureau's narrow core ran at double consensus on the month — and that is the cut feeding core PCE most directly. Richmond's Tom Barkin cited core PCE at 3.7% the same afternoon. The series that set the hike odds and the series the committee argues about are not the same, and on Thursday they disagreed.

The labour data is doing the real work. Initial claims came in at 209,000 against 203,000 expected. Set that beside a July payroll report showing the economy losing 23,000 jobs. It is the combination, not the inflation print, that took September odds down: a committee can look through one soft month on prices, but not through prices cooling and employment contracting together. Which is why today's retail sales matter more than a July number normally would.

The two-year is not saying what it is quoted as saying. The argument circulating — that the front end sits far enough above the funds rate to prove the market prices hikes while the data argues cuts — is directionally right and quantitatively stale. At 4.147% against an effective funds rate of 3.63% the spread is about 52 basis points, not the 75 being repeated, and it narrowed five on Thursday alone. A closing spread is a market resolving the argument, not staging it.

Energy sold off on supply-side arithmetic, not demand data. The IEA and OPEC published the same day and both cut 2026 demand, the IEA revising its balance down by 510,000 barrels a day. That matters twice: it removes the inflation tail that would have made this week's prints dangerous, and it leaves WTI just above the eighty-dollar line a conditional Fed call hinges on — Section 09.

The single names, told once

Applied Materials was the fourth data point and it did not break the pattern. Revenue of $9.12bn, up 24.8%, against $9.00bn expected; earnings of $3.50 against $3.39; an October-quarter guide well clear of the Street. It closed lower and then lost another five percent in the overnight, about 7.6% below Wednesday's close. The company that sells the picks and shovels for the AI buildout said the buildout is accelerating, and the market took a fifteenth of its value out. The stated concern is free cash flow against rising capex. The plainer reading is that the stock had already run 108% this year.

Reddit is today's mechanical event. S&P Dow Jones Indices confirmed after the close that it joins the S&P 500 before Tuesday's open, replacing AvalonBay Communities. The stock closed the regular session up 3.04% and jumped 11.06% to 175.61 overnight. Note the sequencing: the widely-circulated version has that eleven percent happening in Thursday's cash session, and it did not. The index-tracking flow has not traded yet, and it lands in the run-up to monthly expiration.

Workday put up the best day in the index, and Micron added 4.23% on the disclosure that it can return effectively all excess cash once CHIPS Act restrictions lapse in December. Neither is an AI-capex story being sold; the names punished this week are the ones whose thesis is the buildout. Cisco finished down 8.40% a full day after its beat-and-raise, and Tapestry fell 16.49% — worst in the index, a consumer name, going into a consumer print. Meanwhile OpenAI lost its revenue chief, the second senior operating departure in days, at the demand end of the chain Applied Materials supplies. Landing in the same seventy-two hours, they describe a market that has stopped extending credit to the story and started asking each name to pay for itself.

Not blackout, and openly split

No speakers today, no blackout until September 5 — but Thursday made the disagreement public. Cleveland's Beth Hammack, who dissented in favour of a hike last month: “acting now on inflation is really critical.” Richmond's Barkin, hours later: “it remains an open question whether the Fed needs to raise rates.” Same day, no shared premise.

The market has been deciding for them. September hike odds stand at 34.8% against 40.6% a day earlier, 55.0% a week earlier and 75.1% a month earlier — and that month-ago figure carried a 23.9% chance of two hikes, now zero. Cut odds have been zero throughout. Three-quarters of the tightening risk has left the front end in four weeks without the market once entertaining an easing.

Two dates and one condition. Wednesday, August 19 at 2:00 PM brings the July minutes — a meeting held with three dissents in favour of hiking, which makes this an unusually informative set. August 27–29 is Jackson Hole. And the condition came from Jeremy Siegel on Thursday afternoon: if oil stays near eighty dollars, there is no way the Fed raises rates in September. WTI settled at 81.25. On his own terms the condition is met — and the pricing agrees with him.