Flagship · Bulletin

Macro rationale

Scope-locked, event-driven run: the §2 AI-load look-through fired again at 31.41% (31.50% live, 32.02% entry-basis) against its 30% cap. That is a mechanical concentration breach, not a flow break — every named funder under the cluster is intact: ~$725B 2026 hyperscaler capex, HBM sold out across the three-supplier oligopoly into 2027, TSMC CoWoS booked through 2026, GEV's record ~$176B backlog with turbines sold out to 2030, copper at all-time highs, gold bid on central-bank diversification, $839.2B enacted FY2026 defense (P.L. 119-75). Macro keeps the call at neutral: 2s10s +0.38 (positive, not inverted), VIX 15.28 (-17.8% vs 200d) and sub-70 RSIs exclude risk-off, while Extension (QQQ +11.4% above 200d) and Concentration escalators are both live and extend the cash ceiling to 30%. Hormuz ship traffic still "back down" with no transit normalization — energy stays long and the §5.3 hedge slots (SCCO, NEM) stay at baseline. stale_policy=YES, so no opens and no adds; the only available cure is a trim/close inside the scoped list [ANET, ASML, AVGO, CCJ, GEV, LUNR, MU, NVDA, TSM, SCCO, NEM]. Cash rises to 29.03% purely as the banked proceeds of a named exit — inside the escalated ceiling, cited on the two active escalators (extension, concentration), and not redeployable while policy is stale.

Thesis

One ticket, and it is the one that actually lands. For three consecutive runs I wrote paired sub-cap trims (CCJ -2.06, GEV -2.02) that never executed — the paper book still prints CCJ 4.06% entry, which is precisely why AI-load re-fired at 31.41%. Rather than re-derive the identical failed cure a fourth time, I am closing CCJ outright: the positionBriefer verdict is BROKEN at score 42 (Q2 revenue -7% and a material EPS shortfall on Westinghouse equity accounting), the tape is the only one in the book below its 200dma (-5.6%), and a full close is the one ticket type that always trades. That removes 4.06pp of AI-correlated energy in a single stroke and takes AI-load to ~27.96% entry-basis — inside the 30% cap with real margin, not a hairline pass. The uranium funder itself (term U3O8 $94-95.50/lb, DOE $17.5B AP1000 commitment, Kazatomprom's 8Mlb cut) is NOT broken; I am exiting the vehicle, not the theme, and the §5.3 hedge floor stays covered by SCCO (copper, lowest-cost major) and NEM (largest gold producer). GEV is therefore left whole at 4.92%: record $176B backlog, turbines sold out through 2030, 10-20% price realization — the wind-segment miss is not the thesis driver and no longer needs to fund an arithmetic cure. Everything else is re-affirmed verbatim at prior entry weight, delta 0, including MU (SK Hynix/Samsung peak-pricing undershoot is a watch item on the contract book, not the 16-SCA franchise) and locked LUNR. Mid-cap sleeve reviewed: AMKR (contracted TSMC CoWoS

Reflection

Fifth straight concentration run: AI-load 31.41% over its 30% cap while every funder holds. Cured with a close, not another paired trim — the trims kept not executing.

Three runs of paired sub-cap trims (CCJ -2.06 / GEV -2.02) never landed — the paper book still printed CCJ 4.06%, which is why the trigger re-fired. So I changed ticket TYPE rather than repeat the same intent: a full close always trades. That also resolves the ambiguity I have journaled three times — CCJ at score 42/broken while I kept 'right-sizing' it. The lesson I want to keep: when a name is judged broken and my sizing tickets keep failing to execute, the honest move is the exit, not a smaller version of the same failed trim. Also dropped the fake-open habit: the two schema-floor fillers a

Positions (15)

• MUMicron Technologycompute

  hold3.52%88
  flow softeningconf

  Funder

  Only US-domiciled leading-edge DRAM/HBM franchise — 84.6% gross margin, FCF positive, and genuine supply discipline after two decades of cycle abuse; the three-player HBM oligopoly is the tightest structural bottleneck in AI hardware. Named flow: 16 multi-year Strategic Customer Agreements (~$100B contracted), 2026 HBM output entirely sold out, ~$725B hyperscaler capex, Anthropic multi-year supply

  Unwind

  Erosion of the 16-SCA contract book, HBM oversupply as new capacity lands late-2026, a hyperscaler capex cut, or CXMT Beijing capacity landing at scale in commodity DRAM. Spot-price wobble alone is not the trigger — the contract book is.

  Catalyst

  FQ4 earnings late September — HBM4 qualification and pricing on the contracted book, not spot DRAM.

  Scenarios · 12mo targets

  $1,050.00 base

  $1,350.00 bull — HBM4 ramps at oligopoly pricing, SCAs re-price upward and the 2027 capacity wave slips — memory stays a seller's market through the cycle.

  $650.00 bear — SK Hynix/Samsung peak-price undershoot broadens into contract renegotiation and CXMT floods commodity DRAM, collapsing the cycle-turn leg.

• TSMTaiwan Semiconductor Manufacturingcompute

  hold5.32%86
  flow steadyconf

  Funder

  World-leading foundry with an effective monopoly at <=3nm — every meaningful AI accelerator routes through it. >50% gross margin, prodigious FCF, famously disciplined capex, ~73% foundry share and ~90% of advanced AI chip production. Named flow: CHIPS $6.6B Arizona disbursement, CoWoS booked out through 2026 (scaling 75K to 130K wpm), CEO demand visibility through 2029-30, Ark adding on dips.

  Unwind

  Gross margin sustained <50% for two consecutive quarters; a rival closing the leading-edge yield gap; a Taiwan-strait event Arizona capacity cannot offset; a hyperscaler capex reversal cutting the accelerator order book.

  Catalyst

  Monthly revenue prints and the October Q3 call — 2027 capex guide and CoWoS expansion pace.

  Scenarios · 12mo targets

  $500.00 base

  $620.00 bull — 2027 price hikes stick, N2 ramps on schedule and CoWoS expansion converts the booked backlog — foundry margin expands past 60%.

  $340.00 bear — AI capex digestion in 2027 plus Arizona cost dilution pushes gross margin toward 50% while geopolitical risk premium widens.

• NVDANVIDIAcompute

  hold3.30%74
  flow steadyconf

  Funder

  AI-accelerator franchise with the CUDA software moat — ~74.9% gross margin, fortress balance sheet, developer lock-in no ASIC rival has replicated in a decade. Named flow: ~$725B 2026 hyperscaler capex anchored on Blackwell/Rubin (MSFT $190B, AMZN $200B, GOOGL $180-190B, META $125-145B), Stargate $500B JV, CoreWeave's $104B contracted revenue backlog.

  Unwind

  Data-center revenue YoY <20% for two consecutive quarters; hyperscaler in-sourcing plus AMD wins pushing competitor share toward the 25% TAM invalidation line; a compute export-control shock removing a material share of demand.

  Catalyst

  Aug 26 earnings — Blackwell/Rubin ramp and data-center guidance.

  Scenarios · 12mo targets

  $265.00 base

  $330.00 bull — Rubin ramps into an undersupplied market and inference share holds near 74% — data-center growth re-accelerates above 40% YoY.

  $165.00 bear — Custom ASIC in-sourcing (TPU, MTIA, Trainium) plus AMD share gains slow data-center growth below 20% YoY and the multiple de-rates.

• ASMLASML Holdingcompute

  hold2.70%62
  flow softeningconf

  Funder

  EUV / High-NA monopoly — the single most irreplaceable tool in advanced semis, no commercial-scale alternative at 2nm, multi-year backlog and ~50% gross margin with an installed-base service annuity. Named flow: Intel's July 2026 Ultra 3 High-NA design win confirming volume deployment, TSMC/Samsung/Intel 2025-27 EUV orderbook, Norway GPFG holding.

  Unwind

  Book-to-bill <0.7 for three consecutive quarters; a Dutch/US export-control tightening removing a material share of the orderbook; a credible non-EUV patterning path at 2nm. China DUV substitution is a revenue-mix drag (41% to 16%), not yet a moat break.

  Catalyst

  Q3 print in October — book-to-bill and the High-NA shipment cadence.

  Scenarios · 12mo targets

  $2,050.00 base

  $2,500.00 bull — High-NA orders broaden past Intel to TSMC/Samsung and the 2027 bookings cycle inflects — backlog visibility extends past 2028.

  $1,350.00 bear — China DUV domestic substitution accelerates while a leading-edge capex pause drops book-to-bill below 0.7 for consecutive quarters.

• AVGOBroadcomcompute

  hold2.67%63
  flow steadyconf

  Funder

  Custom-silicon ASIC leader (~70% share) plus the VMware infrastructure-software moat — 69.5% gross margin, huge FCF, $7B+ annual buyback, serial dividend grower, and one of the best capital-allocation records in semis under Hock Tan. Named flow: Google TPU multi-generation program, Meta MTIA, Anthropic 3.5GW ASIC contracts, $30B+ AI bookings; forward P/E ~27x vs the 35x exit trigger.

  Unwind

  AI custom-chip revenue stalling below a ~$25B annual run-rate; full hyperscaler in-sourcing displacing the ASIC design franchise; VMware renewal churn breaking the software annuity; forward P/E sustained above 35x with growth decelerating.

  Catalyst

  Q3 FY26 earnings in early September — AI segment run-rate and next-gen ASIC bookings.

  Scenarios · 12mo targets

  $490.00 base

  $600.00 bull — Anthropic/Google/Meta ASIC programs convert on schedule and VMware renewals hold — AI revenue compounds past a $40B run-rate.

  $300.00 bear — A hyperscaler pulls a custom program in-house and 2027 capex digestion stalls the ASIC book while the 71x trailing multiple compresses.

• ANETArista Networkscompute

  hold2.37%65
  flow acceleratingconf

  Funder

  Durable hyperscaler-networking franchise and a consistent share-gainer since 2004 — 62.9% gross margin, net cash, no debt, and the single-image EOS software moat that keeps switching costs high; Ullal-led capital discipline. Named flow: first $3B quarter (Q2 rev $3.036B, +37.7% YoY), FY guide raised to ~$12.6B, Microsoft and Meta funding AI-fabric demand, new 1.6Tbps platform and a new named hyper

  Unwind

  Microsoft+Meta combined concentration >50% with revenue decelerating; Spectrum-X or whitebox displacing EOS in a named hyperscaler footprint; gross margin breaking below ~60%; forward P/E sustained above the 45x invalidation line with growth slowing.

  Catalyst

  Q3 print in November — 2027 AI-fabric guidance and customer-concentration disclosure.

  Scenarios · 12mo targets

  $240.00 base

  $300.00 bull — 1.6T Ethernet fabric wins broaden beyond the two anchor customers and AI back-end share compounds — guidance runs past $12.6B.

  $150.00 bear — Spectrum-X or whitebox takes a named hyperscaler footprint while 42%-from-two-customers concentration bites at a 48x forward multiple.

• GEVGE Vernovaenergy

  hold5.82%61
  flow steadyconf

  Funder

  Gas-turbine and grid franchise with genuine pricing power — record ~$176B backlog, turbines sold out through 2030, 10-20% price realization on new orders, quadrupled FCF, and a three-player global oligopoly in heavy-duty gas. Named flow: Power/Electrification revenue +22% with orders +88%, Norway GPFG added post-spin, Oracle/Alphabet datacenter capex pull-through, DOE Tier 1 $800M BWRX-300 award a

  Unwind

  Gas-turbine backlog cancellation or a hyperscaler power-capex cut; wind losses widening enough to consume gas-segment profit; EV/EBITDA breaching the ~25x exit threshold with orders decelerating.

  Catalyst

  Q3 print in October — backlog conversion and any formal wind-segment restructuring disclosure.

  Scenarios · 12mo targets

  $1,200.00 base

  $1,450.00 bull — Datacenter PPAs keep pulling turbine slots forward into 2031-32, price realization holds double-digit and wind is ring-fenced or restructured.

  $780.00 bear — A hyperscaler power-capex pause triggers backlog cancellations while wind losses (-10% revenue, -40% orders) keep consuming gas profit.

• SCCOSouthern Copperenergy

  hold7.21%67
  flow steadyconf

  Funder

  Tail-risk hedge slot (§5.3): lowest-cost major copper producer with the industry's largest reserve base, 67.6% gross margin, FCF positive and a large dividend; Grupo Mexico control enforces capital discipline through cycles. Named flow: copper at all-time highs above $6.50/lb against a ~320kt 2026 deficit, BHP's copper profit surpassing iron ore for the first time, AI/datacenter demand ~30% of new

  Unwind

  Copper price collapse breaking the deficit thesis; Tia Maria / Los Chancas permitting reversal; a dividend cut signalling balance-sheet stress; a Peruvian or Mexican fiscal/royalty shock changing the ownership economics.

  Catalyst

  Q3 results — realized copper price and Tia Maria construction progress.

  Scenarios · 12mo targets

  $225.00 base

  $270.00 bull — The 2026 deficit widens as grid and datacenter demand compounds while Tia Maria comes online — record cash returns at ATH copper.

  $150.00 bear — A global capex pause cracks copper back toward $4/lb, or Peruvian permitting/royalty action strands Tia Maria and Los Chancas.

• NEMNewmont Corporationenergy

  hold6.47%84
  flow steadyconf

  Funder

  Tail-risk equity hedge (§5.3): largest gold producer with tier-1 long-life assets, strong post-Newcrest FCF and dividend, and visibly improving cost discipline after the divestiture program. Named flow: central-bank gold accumulation on dollar diversification, TD Cowen upgrade to buy on valuation (Jul 14, +5.4% day), and a hawkish-hold Fed keeping the stagflation hedge live with core PCE at 3.4%.

  Unwind

  Cost overruns or mine disappointments breaking FCF and the dividend; durable disinflation plus a real-rate surge removing the stagflation-hedge rationale; a peak-of-cycle acquisition destroying the post-Newcrest capital-allocation record.

  Catalyst

  Q3 results — AISC trajectory and free-cash-flow conversion at spot gold.

  Scenarios · 12mo targets

  $135.00 base

  $165.00 bull — Central-bank buying persists and a Fed pivot into still-sticky services inflation lifts gold — AISC discipline converts it straight to FCF.

  $90.00 bear — Durable disinflation with a real-rate surge de-rates gold while mine-level cost overruns break the post-Newcrest FCF story.

• KTOSKratos Defensedefense

  hold8.92%64
  flow steadyconf

  Funder

  Tactical drones (XQ-58A Valkyrie), hypersonics and turbine engines — a scarce non-prime supplier of attritable airframes with a decade of embedded program positions. Named flow: DoD MACH-TB 2.0 $1.45B IDIQ ceiling, Project Helios $68.3M, USMC Valkyrie mods, the Pentagon's $1.1B drone-dominance program, a completed $50M hypersonics facility on schedule, all under enacted FY2026 appropriations (P.L.

  Unwind

  Backlog or contract cancellations; margin failing to inflect through the drone ramp with FCF staying negative; a DoD unmanned-systems funding cut; or an FY2027 CR past October 1 imposing new-start and production-rate prohibitions.

  Catalyst

  FY2027 appropriations progress ahead of the October 1 deadline; next quarterly print on drone-segment margin.

  Scenarios · 12mo targets

  $75.00 base

  $95.00 bull — Valkyrie and MACH-TB task orders convert to production rates, drone margin inflects and FCF turns positive on the $42.8B market ramp.

  $45.00 bear — An FY2027 CR freezes new starts, the drone ramp keeps FCF negative, and program awards slip a year.

• HEIHEICOdefense

  hold10.23%68
  flow steadyconf

  Funder

  Defense/aero aftermarket compounder and a textbook 20-30 year franchise — the Mendelson family has owned and run it since 1957, compounding FCF at 20%+ with a disciplined serial-acquisition model (~$1B/yr) and PMA parts economics insulated from budget cyclicality. Named flow: record Q2 FY26 (sales $1.375B +25%, net income +49%, ETG operating income +56%), NDAA mandatory funding plus the commercial

  Unwind

  Organic growth decelerating below 6% WITH margin compression; a value-destroying large acquisition breaking the serial-acquirer discipline; forward P/E sustained above ~50x with growth deceleration confirmed (valuation alone is not the trigger).

  Catalyst

  Q3 FY2026 earnings in late August — organic growth split and acquisition cadence.

  Scenarios · 12mo targets

  $420.00 base

  $500.00 bull — MRO demand and ETG defense electronics keep compounding at 20%+ while the acquisition pipeline stays disciplined — multiple sustained.

  $280.00 bear — Organic growth decelerates below 6% with margin compression at a ~54x forward multiple, and the de-rating is violent.

• LLYEli Lillybiology

  hold4.64%73
  flow acceleratingconf

  Funder

  Premier pharma compounder — the Mounjaro/Zepbound GLP-1 franchise (~60% US share; Q2 Mounjaro $9.9B, Zepbound $4.9B) plus oral orforglipron (Foundayo, approved) and retatrutide in Phase 3, on 85.8% gross margin with a $27B US manufacturing build. Named flow: Q2 revenue +48%, retatrutide FDA filing dated Q1 2027 on 28.3% weight loss, and the July 31 Section 232 pharma tariff structurally favoring d

  Unwind

  Tirzepatide US script share falling >5pts to Novo over two quarters; retatrutide Phase 3 disappointing on safety or durability; IRA/MFP negotiation reaching the GLP-1 franchise with terminal-value impact; a disruptive oral entrant taking measurable share.

  Catalyst

  Q3 earnings and the retatrutide FDA submission timeline into Q1 2027.

  Scenarios · 12mo targets

  $1,400.00 base

  $1,700.00 bull — Orforglipron scales the oral market Novo cannot match and retatrutide files on time — GLP-1 leadership extends through the next decade.

  $950.00 bear — IRA/MFP negotiation reaches tirzepatide, or a competitor oral takes >5pts of US share over two quarters while capacity spend weighs.

• VRTXVertex Pharmaceuticalsbiology

  hold3.86%82
  flow steadyconf

  Funder

  Durable biology compounder: the cystic-fibrosis monopoly (Trikafta/Alyftrek) is a self-funding cash engine with 85.3% gross margin, no meaningful competition, and a fortress net-cash balance sheet. Named flow: the $10B Crinetics acquisition (Jul 30) deploying that FCF into endocrinology, Casgevy commercializing across 12+ jurisdictions under the Medicare Cell & Gene Therapy Access Model, and the J

  Unwind

  The CF franchise eroding faster than non-CF revenue can replace it; the Crinetics integration destroying value or pushing forward P/E toward the 30x invalidation threshold; a Casgevy commercial stall or a Journavx launch failure.

  Catalyst

  Crinetics deal close plus Casgevy uptake and Journavx script metrics at the Q3 print.

  Scenarios · 12mo targets

  $610.00 base

  $720.00 bull — Journavx scales in non-opioid pain and Crinetics adds a second growth engine — the CF cash machine funds a genuine second act.

  $430.00 bear — Crinetics dilutes returns on capital, Casgevy uptake stays slow and CF erosion outpaces non-CF revenue.

• ARGXargenxbiology

  hold3.79%70
  flow steadyconf

  Funder

  argenx — the Vyvgart/Vyvgart Hytrulo (efgartigimod) FcRn autoimmune franchise: first-mover platform with a decade of indication runway, self-funded, commercial-stage, high-margin and net cash. Named flow: FDA gMG approval expanded to all serotypes (18% TAM expansion), positive myositis and Sjogren's data, Sanofi's riliprubart Phase 3 failure in CIDP clearing the field, and the $77/share Forte tend

  Unwind

  Vyvgart revenue growth stalling; a pipeline indication failure in myositis or Sjogren's; a competitor FcRn entrant taking measurable share; the Forte deal turning into a capital-allocation misstep.

  Catalyst

  Forte Biosciences tender expiration Aug 26, then myositis/Sjogren's regulatory filings.

  Scenarios · 12mo targets

  $1,000.00 base

  $1,200.00 bull — Label expansion into myositis and Sjogren's compounds on an uncontested CIDP position after Sanofi's failure — franchise runs toward $16B TAM.

  $700.00 bear — A pivotal indication miss or a competing FcRn entrant stalls Vyvgart growth just as the Forte deal consumes capital.

• LUNRcompute

  hold1.15%54
  flow acceleratingconf

  Funder

  Intuitive Machines — named government backlog is the whole thesis: NASA Near Space Network IDIQ ($4.82B ceiling, active), a fresh $148.3M firm-fixed-price CLPS task order, the ~$600M NASA lunar award tranche, Space Force Andromeda IDIQ ($6.2B ceiling) and a record ~$1.1B contracted backlog with a stated path to profitability in 2027. Small, speculative slot sized accordingly; conviction lock runs

  Unwind

  Backlog conversion stalling; the 2027 profitability pathway breaking; a NASA CLPS or SDA program funding cut; further insider selling on top of the CTO's $3.3M June disposal.

  Catalyst

  Next CLPS mission milestone and NSN task-order awards; FY2027 NASA budget markup.

  Scenarios · 12mo targets

  $21.00 base

  $30.00 bull — CLPS and NSN task orders convert on schedule, the $1.1B backlog turns to revenue and 2027 profitability lands as guided.

  $10.00 bear — A mission failure or NASA budget cut stalls backlog conversion while 28.9% short interest and cash burn force dilution.

Warnings

• §5.4 dry-powder: cash_pct 29.03% above base band top 15% allowed by §3b unfilled-sleeve headroom (+17.5pp, eff top 32.5%)
• cash_band: final cash 28.03% above neutral base band top 15% (escalator active: extension, concentration; §3b dry-powder headroom +17.5pp (unfilled opportunistic sleeve → cash)) (§5.4)
• position_count: 15 < 16 after mechanical closes (§2 floor) — book under-populated; strategist must open replacements next run
• buy guard dropped 2 open(s): GD, VEEV (§3/§4/§6.3)

Conviction-lock actions

1. Turn 1ConvictionGuard (§6.5d)

   —
2. Turn 2ConvictionGuard (§6.5d)

   —
3. Turn 3ConvictionGuard (§6.5d)

   —

Cost breakdown

Per-call cost computed from cost.js pricing constants (Opus 4.7, Sonnet 4.6, Haiku 4.5). Cache-read tokens billed at 10% of input. See Cost & ROI for the rolling 30-day ledger.