Market read

Six claims about what kind of market this is. Each one carries the figures it rests on. Two of the mind's words repeat here: to de-rate is to pay less for the same earnings, and dispersion is single stocks pulling apart from each other while the index barely moves.

1. 01

   Mature bull, now rotating

   Everything is up over six months, but leadership has flipped in the last three.

   Value and defensive sectors lead; the AI complex peaked on June 2 and has been quietly distributing under a flat index.

   Breadth is narrowing: equal-weight falls in up weeks and small caps sit below both their 20-day and 50-day averages.

   The figures behind it

   • All indices +9-18% over 6 months
   • SPY bottomed 631.97 on 2026-03-30, +21.7% since
   • 3-month: XLV +13.4%, XLF +13.3%, XLE +10.1% vs QQQ -2.6%, SMH -7.8%
   • SPY/DIA/XLF set 6-month highs in August; QQQ/XLK/SMH peaked 2026-06-02
   • IWM -1.50% below its 20-day, -0.53% below its 50-day
2. 02

   Calm index, distributing insides

   The S&P looks fine on the surface — barely 1% off a record.

   Underneath, the former leaders are 14-25% below their June highs and two stocks are doing most of the lifting.

   Nvidia beat enormously and the stock still gave the pop back within a day: the market believes the demand and de-rates the shares anyway.

   The figures behind it

   • SPY 1.1% below its 8/13 record of 777.88
   • AVGO -25% from June high, SMH -17.7%, TSLA -21.7%, META -16%, GOOGL -14%
   • MSFT +30.5% in a month; +15.5% earnings gap 7/30 (390.54 -> 451.10)
   • NVDA 8/26 beat: revenue +106%, $108B guide, zero-China assumption; +8.7% on 8/27 then -4.6% on 8/28
   • RSP fell in an up week; IWM below its 20d and 50d
3. 03

   Fed flipped from cuts to hikes

   There is a new Fed chair, Kevin Warsh, and at Jackson Hole he put inflation first and scrapped forward guidance — so the market gets less hand-holding into the decision.

   A September rate HIKE went from a fringe idea to roughly a coin flip in a single day.

   The data underneath is contradictory: surveys are hot, hard data is soft. That is a stagflation-shaped problem, and it is the hardest kind for a central bank to answer.

   The figures behind it

   • Chair Kevin Warsh since 2026-05-22; Jackson Hole keynote Fri 2026-08-28
   • Sept-16 FOMC hike odds ~55-60% (FedWatch 57%, Kalshi 48%), up from ~35% a day earlier
   • Fed funds target 3.50-3.75%
   • July NFP -23k with -103k of prior revisions
   • ISM manufacturing 55.6 (4-year high); CPI 3.4% headline / 2.5% core; GDP 1.5%; new home sales -10.5%
   • July FOMC held 9-3 with three hike dissents
4. 04

   Fiscal dominance is the undertow

   Long-term government borrowing costs hit their highest since 2007, and the Treasury has had to step in and buy back its own long bonds to hold them down.

   The debt pile crossed $40 trillion, and the AI build-out is borrowing into the same bond market.

   When rates shock the tape, it hits semis, small caps and gold far harder than it hits the S&P — Friday was a live sample of exactly that.

   The figures behind it

   • 30-year Treasury yield hit 5.33%, highest since 2007
   • Treasury doubled long-end buybacks ($2bn -> $4bn per operation), starting 2026-09-09, running to 2026-11-04
   • US debt crossed $40T
   • $969B of undiscounted hyperscaler data-centre lease commitments at end-2025 (~113% of their on-balance-sheet debt)
   • Friday 8/28: SMH -3.5%, IWM -1.4%, GLD -3.2%, SPY only -0.2%
5. 05

   Vol priced for none of this

   Option prices say the next month is going to be quiet, right as the calendar gets its most binary in months.

   The cushion for anyone selling that calm is thin — implied vol is barely above what the market has actually been doing.

   What the vol market IS paying up for is single-stock AI dispersion, not index risk. Upside index calls are the cheapest thing on the board.

   The figures behind it

   • VIX 14.43, a YTD low
   • SPY 1-month ATM IV 11.62% (parity-clean), 1-week 9.75%
   • SPY 20-day realised vol 10.40%; 10-day 7.83% — a thin +1.2 point premium
   • CBOE equity put/call 9-day average at the 13th percentile
   • QQQ/SPY 1-month vol ratio 1.47 vs a normal 1.25-1.35
   • Sept-30 787 call at ~10.29% IV — cheapest optionality on the surface
6. 06

   Positioning stretched against itself

   Professional investors are as long as they have been in years and holding almost no cash — historically a contrarian warning.

   Retail investors are the opposite: unusually bearish.

   And the obvious 'crowded AI trade unwind' has already partly happened, which argues against the lazy version of the bear case.

   The figures behind it

   • BofA FMS cash 3.5%, below the 4.0% sell-trigger
   • Net +56% overweight equities, highest since Nov 2021
   • Record share expecting a 'no landing' outcome (56%)
   • AAII bears 44.4%
   • 'Long global semis' as most-crowded trade fell 82% (July) -> 53% (August)

Every event the mind logged, in date order, from today onward. Its thesis is a bet on this stretch, so this is the calendar it is trading against. Times are US Eastern; an entry with no time runs all day. AMC beside an earnings date means after the market closes.

Four research agents, one question each. Their findings are the raw material the six-point regime read was distilled from.

These are not positions. The mind holds nothing. They are names it flagged as worth a session of their own, with the reason and the moment it would look.

• AVGO (Broadcom)

  The highest-information event of the week. The market already assumes a beat (prediction markets ~95%), so the AI guide and the margin are the actual event. MRVL fell 10.3% after its own report — that is the bearish read-through. Down ~25% from its June high while the S&P sits near a record.

  When: Earnings Wed 2026-09-02 after the close. Decide Monday or Tuesday with live implied vols.

• BF.B (Brown-Forman)

  A ~16% implied move priced on a consumer-staples whiskey company — the most obviously expensive event of the week. A candidate study in whether rich vol is actually rich.

  When: Earnings Wed 2026-09-02 before the open.

• PYPL (PayPal)

  A ~$50B takeover broke and the stock fell 12.7% on Friday; David Einhorn is newly long. Either a value trap with the bid removed, or a forced-seller overshoot with a credible activist anchor — a study in which.

  When: No scheduled catalyst; an ongoing situation, needs its own session.

• XLF calls (financials)

  The cleanest uptrend on the board (+13.3% over three months, essentially at its six-month high) with cheap implied vol. But the rate decision cuts both ways for banks, so the trade is blocked until the uncertainty clears.

  When: After the Sept 16 FOMC.

• TLT (long Treasuries)

  Genuinely two-sided: the stressed 30-year auction on Sept 10 pushes one way, the Treasury's upsized buybacks push the other. Needs research before it is tradeable.

  When: Around the Sept 10 auction; research first.

The thesis page keeps two of the mind's own words: Goldilocks data means growth steady and inflation cool enough to suit both sides, and a melt-up is a fast rally driven by chasing rather than by news.