3. Three Things

1. Oracle booked $664 billion of backlog, grew cloud infrastructure revenue 121%, and fell 1.74% on 3.27× volume — while $DELL (+12.0%) and $HPE (+12.4%) were the S&P 500's two best stocks on the same print.
2. August inflation ran hot and the bond complex sits at capitulation-adjacent readings — the 7–10 year Treasury fund at RSI 26.4, high yield at 28.3, all five bond funds death-crossed — yet the S&P 500 fund rose 0.8% and volatility fell 11.2% to 15.84.
3. $SWKS closed at RSI 83.4 after a 19.4% week on reported progress toward the $22 billion Qorvo merger and a quarterly beat — the single most extended reading anywhere on the board.

4. The Big Picture

Friday was the day the market decided it could live with a rate hike.

Act 1 — The bond market capitulated and equities ignored it. The five bond funds on the macro dashboard all sit at or near capitulation: 7–10 year Treasuries at RSI 26.4 (down 12.5 points in a week), the total-bond fund at 28.7, high yield at 28.3, investment grade at 29.8, the 20-year at 35.2. Every one is death-crossed, negative on the year, and below its 200-day average — the most damaged corner of the entire board, and it got worse this week rather than stabilizing. Housing tracked it down: the homebuilder funds fell 4.6% each on the week, now 7.9% and 9.0% below their 200-day averages and down 17.1% and 22.8% on the year. Against that, the four broad equity benchmarks cooled on RSI but held every long-horizon field — the S&P 500 fund +16.22% on the year and 7% above its 200-day, the Nasdaq 100 fund +22.39% and 8.3% above. The oversold band on that dashboard more than doubled, from 8 names to 18, while 53 of 56 names cooled. That is a deep pullback with rotten internals, not a breakdown.

Act 2 — The buildout got paid, the landlord didn't. Oracle proved compute demand is contracted and deliverable, and the market paid the supply chain for it. Beyond Dell and HPE, the lanes that make the physical plant closed as coordinated blocks: transformers and grid interconnect +3.6% with all 8 up, power distribution +3.6% with all 10, high-speed connectors +3.6% with all 10, optical transceivers +3.6% with 22 of 24. Thirty-three of 124 lanes closed at least 90% green; only one closed 90% red. Underneath, the rotation map's accelerating cohort is entirely capital-equipment: semiconductor test and burn-in, data-center networking silicon, optical transceivers, contract electronics, high-speed cabling. The equipment book is the cleanest long-horizon read on the desk — all 14 names positive on both the trailing year and excess return, no exceptions, with $FORM (+9.0 RSI), $KLIC (+8.4) and $AMKR (+8.2) leading the advance and $MKSI crossing back above its 200-day.

Act 3 — The wreckage is in the consumer and the fuel cycle. The nuclear fuel lane fell 4.1% at the median with 3 of 19 up: $SMR -15.7% on 2.73× volume with no cause established on its receipt, $OKLO -9.2% on 2.47× volume the day it filed a prospectus supplement, $LEU -8.2% and now 27.1% below its 200-day. Retail and defensives are worse and quieter: $TJX at RSI 17.8, a third consecutive week deeper into capitulation with nothing beyond an August earnings report on its receipt; $HRL at 23.2 and -19.1% on the year; $SYK at 27.8 and -29.1%. Defense keeps selling into a live Middle East war — $LHX at RSI 24.9 in a collapse regime, 21.7% below its 200-day and negative on the year; the aerospace fund at 28.5. There is no war premium anywhere in the defense complex, and the one name with a genuine catalyst, $AVAV, gained 4.4 RSI points on a 37% backlog increase while still sitting 30.1% below its own 200-day.

5. Focus List

$MU is the level to write down. At $975.26 it sits under the $989.96 flag ceiling for a second session, which keeps its buy condition at wait regardless of a 56.8% cushion above the long average. The week was -4.1% against a month that is still +7.0%; nothing about the setup is broken, the trigger simply isn't back.

$SHOP changed category this week, not just price. An 11.2% weekly decline took it from a pullback regime to a downtrend and from above its 200-day to 1.7% below, with the one-year return at -11.2% and 27.4 points of negative excess return. The news pull ran Friday evening and returned quote pages only — no article behind the move. This is now a trend under test, not a dip in a winner.

$NET is the opposite problem. It posted the list's best week at +9.9%, flipped back to an uptrend regime, and sits 36.5% above its long average — on a freshly-pulled receipt containing nothing but portal listing pages. A move this size with no attribution is a position to respect and not a thesis to add to.

$PL is the horizon lesson in miniature. Up 70.1% on the year and 39.3% below its 200-day, at RSI 23.5 after a 32.9% month. The one-year number says this was a monster; the long average says the monster is being dismantled. Interest begins at a higher low, not at a round number.

$GOOGL's cushion is now the story. It closed +1.8% on the day and is up 40.8% on the year, but only 0.5% above its 200-day after basing out of a pullback. A single bad session takes it below the line that separates its two readings.

6. Sector Scorecard

7. Market Vibe

The tape is bifurcated along a single axis, and it is not AI-versus-everything. It is who pays for the buildout. Names that sell picks and shovels to the buildout rose as coordinated blocks on Friday — racks, switches, power semiconductors, thermal, transformers, connectors, optics, all with 90%+ of members green. Names that fund the buildout with their own balance sheet had a much more mixed session: the GPU-cloud independents rose, but from deep holes (Core Scientific 9.2% below its 200-day, Applied Digital 18.2% below, Cipher 8.6% below), and Oracle, the best-capitalized operator in the whole category, finished last of its own cohort after a quarter almost everyone agrees was operationally excellent.

What doesn't fit the narrative is the rate market's non-reaction in equities. Inflation ran hot, hike odds jumped 38 points in a week, Treasury yields sat near multiyear highs, and the fear gauge fell 11.2% to a complacent 15.84 while the breadth gauge flipped to short-term fragile. The rotation map still shows roughly three lanes flowing out for every one flowing in, with leadership concentrated in crypto, memory and Iran-war read-throughs rather than broadening. Cheap volatility on thinning participation is complacency without confirmation — that specific divergence went on the board Friday as a fresh regime observation, not an entry signal.

The second thing that doesn't fit: gold and silver. In a week with a shut Saudi pipeline, Houthi advances toward the Bab el-Mandeb chokepoint, record $6 diesel, and an active Middle East war, the gold fund fell 1.97% and sits 4.1% below its own 200-day, and silver fell 2.84% to 11.4% below. Metals are trading the real rate, not the geopolitics. So is defense, which keeps selling. If you were positioned for a war bid, the tape has now told you three weeks running that you are in the wrong instrument — the war is expressing itself in crude, diesel and shipping, not in havens or primes.

8. Scan Dashboard — Every Manifest Scan at a Glance

6a. The Dashboard Table

6b. Expanded Dispatches

Semiconductor Equipment. Every one of fourteen names carries a positive trailing year and positive excess return — the only book on the desk with no exceptions, ranging from $ACLS at +15.1 points of excess return to $FORM at +259.8. Nine gained RSI this cycle, led by $FORM (+9.0 to 53.9), $KLIC (+8.4 to 48.6) and $AMKR (+8.2 to 49.5), and the below-200-day group shrank from five names to four as $MKSI crossed back above. Four names remain under their long average — $CAMT, $AMKR, $ACLS and a research-only European listing — and all four still carry strongly positive years. This is the textbook deep-pullback-in-an-intact-trend book, and the rotation map has the lane turning: 30-day relative strength still -11.8 against 7-day at +2.1.

Momentum Board. 443 names across 143 watchlists, down from 448 — rotation, not a breadth collapse. The surprise is what's leading: the parabolic tier carries nine biotech names ($MRNA, $TXG, $MMED, $IOVA, $ABCL, $CDNA, $TWST, $AMLX, $CORT) and five refiners ($DK, $DINO, $MPC, $CVI, $VLO), with AI infrastructure only the third cluster. Defense has no two-name cluster in the top three tiers at all. Twenty-seven names print RSI ≥70, concentrated almost entirely in refiners and tankers — due a pullback to the 20-day, not a fresh add. The steady-climber tier, which is where trend-hold entries come from, added thirteen names led by $AAPL and an energy-major run ($CVX, $OXY, $SHEL, $BP, $OKE).

Defense Contractors. The sector's selloff didn't reverse, it ground on. $LHX fell into a collapse regime at RSI 24.9, 21.7% below its 200-day and now -12.5% on the year, worse than -6.06% a week ago. $LEU dropped 8.18% in a single session and 12.41% on the week, taking its year to -32.1% and its gap below the long average to -27.1%. No defense-specific receipt is fresh enough to explain either: the pulls for the primes range from late April to mid-August. What lines up instead is the dated, market-wide rate shock — which is a read-through, not a company cause, and is stated as such.

Biotech. The cohort turned over completely: two cycles ago the overbought club held nine names, now the entire cut sits below RSI 62 and the oversold band went from empty to three names deep. $RARE is at 20.2 and -54.2% on the year with a receipt that predates the slide; $AMGN collapsed 31 RSI points on a 13.7% week while staying 5.0% above its 200-day with a +33.5% year — oversold RSI, intact long trend, and no cause established on a receipt pulled Friday evening. The fund tells the opposite story to the single names: +63.22% on the year and 14.7% above its long average.

Optical Supply Chain. 28 of 36 names gained RSI, and only two — $FN and a small-cap laser name — are broken on every horizon. $CIEN rejoined the recovery at +9.4 RSI points behind a fiscal third quarter its own receipt confirms: $1.67 billion of revenue up 37% year over year, adjusted earnings per share of $2.11 up 215%, and full-year revenue guidance raised to $6.42 billion on AI-driven demand — and it is still 8.1% below its 200-day, so that is a recovering print, not a resolved trend. One row is untrustworthy: a Japanese OTC mirror printed a price and zero volume byte-identical to last week's, a full week frozen; it is reported as unconfirmed, not flat.

9. The Front Page — Deep Research Teasers

7a. Perspective Dispatches

• AI power: rack and chip conversion — INTACT, bounded. Grid-equipment demand supports the delivery layer; rack, board and die beneficiaries still need their own design-win and margin evidence. Friday gave the cleanest tape confirmation yet: power semiconductors +4.2% with all 12 members up, power supplies and shelves +4.1% with all 7, and $VICR +11.2% on a receipt that carries no cause.
• AI optics: bandwidth becomes supplier earnings — INTACT, provisional. Supplier allocation and margin evidence is still required before promotion. $COHR is up 8.3% on the week against -14.1% on the month and now carries four of five convergence bells, the loudest reading on the board; $GLW is +7.8% on the week with an 8-K filed Friday.
• AI memory: pricing discipline through the supply ramp — INTACT. Producer pricing, shipments and customer agreements sustain it; inventory and capacity additions remain the decisive falsifiers. The week added a demand-side data point from outside the tape: CNBC reported September 11 that sub-$100 smartphones are becoming a rarity as AI memory prices soar. $MU is +547.71% on the year and $SNDK +1,837.5%, both in pullback regimes.
• Nearline storage: density and supplier discipline — INTACT, provisional. $STX fell 3.7% on the day to $830.17 and is -5.5% on the month while still +321.81% on the year and 37.0% above its 200-day; $WDC is -3.0% and +365.09%. Allocation and competitor-capacity evidence is still the open item.
• Nuclear fuel: contracting reaches enrichment and supply — INTACT, bounded. This was the week's worst lane: -4.1% at the median with 3 of 19 up, $LEU 27.1% below its long average, $SMR -15.7% on 2.73× volume and $OKLO -9.2% on 2.47×. Future agreements support the demand branch; fuel and enrichment economics still need their own production evidence.
• Biotech funding: conversion into tools and CRO orders — INTACT, untriggered. The shared momentum condition remains incomplete. The tools names are candidates rather than trends — $RGEN -1.4%, $CRL -3.5% and $RVTY -4.5% on the week — and funding activity still has to connect to downstream orders before promotion.
• Energy disruption: physical flows and exporter economics — INTACT, conditional. Price strength keeps the scenario live: the oil fund +112.19% on the year and 41.9% above its long average, the Brent fund +105.9%. The physical leg got dated evidence this week — Saudi Arabia's East–West pipeline shut after Houthi attacks, and US diesel topping $6 a gallon (CNBC, September 11) — but company-level transmission still requires separate verification.
• Drone defense: contracts become profitable deliveries — WOUNDED, market expression. 14 of 21 names now carry negative excess return, up from about half. $AVAV's genuine beat-and-raise moved RSI 4.4 points inside a chart 30.1% below its 200-day. The demand source is plausible; the market expression is not working.
• AI spending: the supplier earnings reversal case — INTACT, risk lens. Oracle's quarter challenged the reversal case on the leg it rests on — capital expenditure of $28.499 billion, fiscal 2027 guided above fiscal 2026's $55.663 billion, more than $30 billion of new contracts — and supplied real evidence on its financing channel instead: the $20 billion equity program fully drawn, $11.363 billion of customer prepayments, a fifth negative free-cash-flow quarter, and $125.0 billion of senior notes marked at $105.7 billion. Cost of money is not a disclosed spending reduction; the case's support condition is not met.
• AI power: generation and grid equipment — INTACT. The interconnection leg is still at queue position rather than delivered output: Oracle's filing quantifies $34.150 billion of unconditional purchase and power obligations and $288 billion of not-yet-commenced data-center leases, but no contracted or energized megawatt figure anywhere. Commitment is documented; energization cannot be verified from the filings.

Every perspective quoted here carries a September 11 data stamp, and every number above is taken from the settled September 11 close rather than the perspective body.

7b. Momentum Watch

Biotech and refining are the week's two dominant clusters, not AI — nine biotech names and five refiners in the parabolic tier alone. AI infrastructure is a real but third cluster, and its members are the hardware and memory names ($DELL, $HPE, $MU, $SNDK, $MRVL, $NBIS, $ANET, $NET, $SNOW, $CIEN) rather than the model or cloud layers. Twenty-seven names print RSI ≥70 and they concentrate in refiners and tankers — that is where the pullback risk sits, not in the semis.

7c. Deep Dive Spotlight

Oracle — "Oracle Emptied Its $20 Billion Equity Program in a Single Quarter." The demand question is answered; the financing question is now the entire thesis, and Friday's filing made it harder. Backlog of $664 billion against $455 billion a year ago, infrastructure revenue +121%, 850 megawatts and 300,000-plus GPUs delivered in one quarter — set against $28.499 billion of capital expenditure on $23.103 billion of operating cash flow, the entire at-the-market authorization consumed for 141 million shares, half of record operating cash flow arriving as customer prepayment, and $288 billion of leases not yet commenced. Verdict: research-watch, conviction Low. The constructive condition is a base that reclaims the $167.64 long average on above-average volume; the destructive one is a loss of $140.68 with Thursday and Friday's volume pattern repeating.

Nebius — "Nebius Proved the Economics. NVIDIA's 9.3% Stake Makes the Demand Question Harder." Real growth, real contract economics, real operating leverage; the investment improves the funding path while making independent end-user demand the next essential proof. Verdict: research-watch, conviction Medium. It closed -1.56% at $224.55 on the day the rest of its lane rose, and is +151.77% on the year and 42.4% above its 200-day — the healthiest long structure in the category by a wide margin.

Vicor — "Vicor Finally Has Orders, Not Just Ideas. Now It Has to Convert Them." The business inflection is real; the stock setup is not clean yet. Verdict: research-watch, conviction Medium. It closed +11.2% at $197.91 on a receipt with no cause behind the move, and sits 3.2% below its 200-day with a +288.4% year — an orders story trading under its own long average.

Plus 50 more tracked-company profiles refreshed this cycle.

7d. Sector Rotation Radar

10. Key Signals

Rates and the Fed. August consumer prices rose more than expected, energy-led per CNBC's own breakdown, and the reaction firmed rather than faded: Bloomberg reported Treasuries closing a bruising week with hike bets firmed and the dollar wavering, and former New York Fed president Bill Dudley said publicly he'd be shocked by anything other than 25 basis points. Crowd pricing on a September hike stands at 81%, up 38 points on the week; odds of a hold fell 34 points to 21%. Ten-year yields were last observed at 4.95%, the two-year at 4.56%, the curve at +0.33%, high-yield spreads at 2.70%, and the real ten-year at 2.55%. The tripwire remains 5.00% on the ten-year or 3.50 points of high-yield spread on two consecutive readings.

Volatility and breadth — the week's named divergence. Volatility fell 11.2% to 15.84 and back under the complacency line in the same session the desk's breadth gauge flipped from deteriorating to short-term fragile, with the rotation map still showing roughly three lanes on net outflow for every one on inflow. Fear pricing falling alongside thinning participation is complacency without confirmation. The trigger that would resolve it constructively: volatility holding under 16 for two straight weeks while breadth returns to improving. The invalidation: volatility back above 20, or breadth reverting to deteriorating while volatility stays compressed. Single observation, no confirmations yet — it is a regime read, not an entry.

Gold and commodities. The energy complex is the only leadership that strengthened: crude's fund +9.12% on the week to +112.19% on the year and 41.9% above its long average, the Brent fund +9.4% on the week, the E&P fund at RSI 70.2 in a breakout regime. Both crude funds fell on Friday itself. The metals went the other way — gold -1.97% on the week and 4.1% below its 200-day, silver -2.84% and 11.4% below, both deepening despite a shut Saudi pipeline and Houthi advances on the Bab el-Mandeb. Copper's fund cooled 2.35% on the week while holding +67.04% on the year. Metals are trading real rates; crude is trading physical supply.

Insider buying. The buying side narrowed from five names to three — $ASAN (+$8.7M), $INTC (+$2.4M), $TSM (+$1.5M, down from +$1.8M) — with $TSLA and $CRSP dropping off entirely, which the pull cannot resolve as either a reversal or a threshold effect. No name in that 59-stock universe has closed under RSI 30 for six straight cycles, so the oversold-plus-buying combination is empty again. Selling stays broad at 29 names, led by $LLY (-$3.0B), $PLTR (-$2.4B), $AVGO (-$1.6B) and $PANW (-$1.4B) — but note that $AMD, $MRVL and $DOCN carry heavy selling against strongly positive multi-year trends, while $NFLX carries it against genuine chart damage (-35.69% on the year, 10.3% below its 200-day). That last one is the only case where selling and the chart agree.

Crypto as a risk gauge. Bitcoin at $77,319.41 golden-crossed with +21.95% behind it on the month, and ether, solana and chainlink all did the same. Every coin is still negative on the year, with bitcoin's -33.06% the shallowest deficit in the tier by a wide margin and no alt-season rotation visible on the week. The equity layer contradicted the coins: $MSTR crossed back below its 200-day (-5.1% from +2.7%) and $HOOD's one-year return flipped negative again, both with no sourced catalyst on their receipts. Risk appetite is in the coins, not in the vehicles that lever them.

Regime transitions. The newest batch carries 54 state changes against Thursday, 45 of them confirmed, with no entries and no exits. The shape is defensive: uptrend → pullback is the largest bucket at 12 names, followed by downtrend → collapse at 8 and pullback → downtrend at 6, against only 3 pullback → uptrend and 3 basing → uptrend. On research-focus names: $NET moved pullback → uptrend, $GOOGL pullback → basing, $SHOP pullback → downtrend, $IREN collapse → downtrend. Board-wide the states stand at 413 pullback, 253 uptrend, 166 collapse, 149 basing, 123 downtrend, 11 breakout, 5 parabolic.

11. The Wild & Whacky

• Oracle proved the demand case and finished last in its own cohort. Backlog up $209 billion year over year, infrastructure revenue up 121%, 300,000 GPUs delivered — and the stock closed -1.74% on 3.27× volume behind nine GPU-cloud independents, six of which were green. The suppliers took the whole payout.
• A momentum screen surfaced a takeout. $ATAI sits in the momentum tier at 55.8% above its 200-day and +93.4% on three months, last traded September 10 at $7.35, and did not print a Friday bar — because its own 8-K filed September 11 describes a sale to Eli Lilly at $6.75 cash plus a contingent value right. The screen is measuring deal arbitrage, not momentum.
• $TJX has now spent three consecutive weeks deepening into capitulation at RSI 17.8 — the lowest print anywhere on the desk — with a receipt pulled Friday evening that contains a quote page, a portal page, and an August 19 earnings report. Nothing explains a 17.4% monthly decline in a discount retailer during an inflation scare.
• $PHGE rose 30.3% on the week on 16.71× average volume with a receipt whose newest substantive item is from April — and it is down 97.9% on the year and 91.8% below its 200-day. A 98%-down microcap bouncing on sixteen times normal volume is a lottery ticket, not a turnaround.
• Defense sold off into a shooting war for the third straight week. $LHX at RSI 24.9 in collapse, the aerospace fund at 28.5, and only $RTX still above its 200-day among the primes — while diesel hit a record $6 a gallon and Saudi Arabia shut a pipeline.
• $SWKS at RSI 83.4 is the most extended reading on the board, and it has a real reason — reported progress on the $22 billion Qorvo merger plus a quarterly beat, per Yahoo Finance and TIKR on September 11. Last week the same move had no receipt at all; the pull caught up.
• Two inventory builds worth watching. $PWR's inventory grew 64% year over year against 41% revenue growth in its June quarter, and $VRT's grew 78% against 24%. In a buildout, inventory ahead of revenue is either pre-positioning or the first sign of an order pause; it resolves at the next print, not before.
• An AI forecaster is trading the thesis in options. CNBC's David Faber reported September 11 that Leopold Aschenbrenner's Situational Awareness has been buying options positions in $AMD, $BE and $CRWV. That is sourced positioning, not evidence — but all three are financing-side names, which is the exact axis this week sorted on.

12. Paper Trade Report Card

Eleven paper books sat their sessions on the Friday close. Ten passed or held on their own written rules; the leveraged-speculation book trimmed half of its oil-explorers position under its +15% partial-profit rule (filled at $195.72, +15.01%). Correction, later Friday night: that book's risk gate had been dark since September 5 because of a phantom, not a hole — its June 7 SOXL exit had been filled at a two-session-stale close, booking a +25% target that never printed on the day it was dated. That fill, and three like it in two other books from the same June 7 run, were re-booked at the June 5 settled close. The speculation book drops to second on the leaderboard ($109,270.76, +9.27%) behind the wealth-advisor book ($109,307.92, +9.31%), and its gate is clear. Two typed buy proposals (a $1,000 DuPont entry in the mechanical-signals book and a $2,971 crypto-industry fund entry in the swing book) were submitted after the close and fill at Monday's canonical noon print, not at a price invented tonight. At session time the leveraged-speculation book's risk breaker read as held since its September 5 fill; the correction above is why, and it no longer holds.

13. What I'd Tell a Friend

The market paid the supply chain for Oracle's quarter and charged Oracle for it. That is the trade to sit with: own the equipment, not the balance sheet doing the borrowing. But Friday's prints on the winners were 12% single-day moves, which is the worst possible entry for a trend-hold book. Nothing below is a chase.

1. $COHR — in zone, four of five bells, the loudest signal on the board. It closed $305.37, above the $299.22 reclaim its watch names, +8.3% on the week against -14.1% on the month, 9.2% above its 200-day with a +195.0% year. Entry condition: a hold above $299.22 with the week's volume pattern intact. Invalidation: a close back under $299.22, which returns it to wait.
2. $GLW — the trigger is already cleared and the structure is clean. It closed $166.40 against a $154.59 level, +7.8% on the week, 14.4% above its 200-day with a +120.5% year, and filed an 8-K Friday. Entry condition: a first pullback that holds $154.59. Invalidation: a close below it.
3. $MRVL — cleared $224.58, closed $236.10, +5.6% on the week with a +254.6% year and 52.5% above its long average, still 28.4% off its own high. Entry condition: the $224.58 level holds on a retest. Invalidation: losing it.
4. The semiconductor equipment book is the cleanest long-horizon group on the desk — all fourteen names positive on both the trailing year and excess return. $TER ($379.72, +228.6% on the year, 18.4% above its 200-day) and $FORM ($114.95, +276.0%, 11.3% above) are the two with both the trend and the cushion; $AMKR (+103.8% on the year but 8.4% below its 200-day) is the trend-under-test version and needs the long average reclaimed first. Entry condition for the group: a reclaim of the 50-day on above-average volume after a higher low.
5. $MU is not a buy until $989.96 is back. It closed $975.26 for a second session under the flag ceiling, with a +547.7% year and 56.8% of cushion. The number is the number; the setup is intact and the trigger is not.

What not to do today: chase energy. The refiners are at RSI 74–83, the E&P fund at 70.2 and the crude fund at 71.9 with 41.9% of cushion above its long average. That trade is made. Same for $DELL and $HPE after 12% days — the thesis got confirmed, the entry got worse.

The one thing to watch: the Federal Reserve's decision is scheduled for Wednesday, September 16, with crowd pricing at 81% for a hike. Equities spent Friday deciding they could absorb it while the bond market sat at capitulation readings. If the ten-year prints 5.00%, or high-yield spreads reach 3.50 points on two consecutive readings against today's 2.70%, the levels above stop being entries and start being exits.

14. Active Perspectives

15. Scan Summary

Thirty-one scans completed on the September 11 settled close: Market Pulse, AI Sector, AI Infrastructure, Biotech, NVIDIA Ecosystem, Defensives, Healthcare, Consumer, Retail, Cloud ETFs, Cybersecurity, EV & Clean Energy, ETF Universe, Insider Activity, Bargain Bin, Tech Insider Buys, Crypto, Macro & Commodities, Defense Contractors, Geopolitical Risk, Optical Supply Chain, Drone & Defense, Supply-Chain Traces, Semiconductor Equipment, Food Security, Cultural Signals, Airlines, Chemicals, Travel & Leisure, Momentum Board and Turnaround Watch. Momentum Discovery ran alongside as an optional screen.

Ten perspectives carry a September 11 data stamp. Fifty-three tracked-company profiles were refreshed, led by the Oracle dive. Eleven candidate themes were re-cut. Coverage reached 2,579 rows at the September 11 close across 1,139 unique symbols, with non-derivative breadth of 753 up, 359 down and 14 flat across 1,126 names at a median one-day change of +0.6%.

Data source: the desk's settled September 11 close. Event attribution comes from company filings, wire and publication coverage dated September 8–11, and where no such receipt exists the move is reported as having no established cause rather than given one. The prior issue is the September 10 daily brief, which set the cohort test this issue resolves.