Flagship · Bulletin

Macro rationale

Digest holds NEUTRAL and I agree. The stagflation-defensive frame is intact: July NFP -23k (first negative of cycle, -103k revisions), headline CPI 3.4% with energy +14.7% YoY, Fed pinned at 3.50-3.75% with three HIKE dissents under Warsh, 10y real 2.41%. Against that, credit and vol are complacent — HY OAS 2.73 near its SMA, VIX ~15.8, 2s10s +0.38. No acute stress signal fired, so hedges stay baseline. The only active escalator is EXTENSION (QQQ +8.71% above 200d), which lifts the enforced ceiling to 25% but does not change the regime. Internals corroborate a rotation, not a break: XLV +6.64pp, XLE +5.83pp, XBI +5.17pp lead over 1m while XLI and XLU lag; equal-weight is +1.64pp vs SPY, so breadth is broadening rather than narrowing — the opposite of a top-of-cycle tape. Pillar lens: defense +0.86pp, biology +5.17pp, compute -0.76pp, energy -0.23pp. Cash: carried entry cash of 28.09% is ABOVE the 25% escalated ceiling and must be cured. I bring it to 23.0% — inside the escalated band, comfortably above the 17% working reserve — with the named cause being (a) the fail-closed ceiling cure and (b) funding the first two names of the §3a foundation sleeve, which the sleeve mirror flags as empty against a 10-20% band and which the Aug-14 brief explicitly wants overweight in a late-cycle stagflation with a pinned Fed. This is not a re-expression of the same regime read week to week; it is a mechanical cure plus a named deployment.

Thesis

Zero thesis-name tickets. The fire is three 13F-exit headlines — Druckenmiller out of AVGO and MU, Loeb out of AVGO, Laffont down 88% in NVDA since Q1'23 — and not one of them is a funder I ever cited. My AVGO thesis rests on GOOG TPU multi-gen, META MTIA and Anthropic ASIC programs plus VMware renewal economics; my MU thesis rests on the 16 Strategic Customer Agreements worth ~$100B of contracted HBM and a three-player oligopoly with 2026 supply fully pre-sold; my NVDA thesis rests on ~$725B of 2026 hyperscaler capex and CUDA lock-in. A famous manager taking profit after a 231-300% run is a valuation opinion, not a capex cut, not a contract cancellation, not a margin break. Notably the same tape shows Laffont ADDING to MU and Druckenmiller accumulating NVDA — the "smart money" signal is incoherent in both directions, which is exactly why it is not in my unwind conditions. Haiku high-flags run 76% at 7d but only 53% at 30d with negative average alpha; I weight them as noise on a multi-year hold. KTOS -7.2% today and HEI -2.9% are price, not news: KTOS finished its hypersonic test facility early and HEI just printed FSG $929M / ETG $460M beats with ETG operating income +56% — HEI's >50x-P/E invalidation requires DECELERATING growth alongside, and growth is accelerating. So the whole thesis core is copied verbatim at entry weight, delta 0. The two real decisions: cure carried cash of 28.09% back inside the 25% escalated ceiling, and start the empty §3a foundation sleeve with 2.

Reflection

Three billionaire-exit headlines (Druckenmiller out of AVGO/MU, Laffont down 88% in NVDA) and zero funder breaks. Sentiment flow, not capital flow. No thesis tickets.

The signal I keep getting fired on is 13F sentiment, and it is incoherent: Druckenmiller exits MU while Laffont adds; Laffont cuts NVDA while Druckenmiller accumulates. Two famous managers taking opposite sides of the same name is definitionally not information. My unwind conditions name contracts and capex lines for a reason — I should probably say so explicitly rather than re-litigating each headline. Separate thing nagging me: the foundation sleeve has been empty for weeks while I journal about compute drift. The brief has wanted quality ballast overweight since Aug 14 and I have not built

Positions (14)

• MUMicron Technologycompute

  hold3.78%87
  flow acceleratingconf

  Funder

  Only US-domiciled leading-edge DRAM/HBM franchise: 84.6% gross margin, FCF positive, and real supply discipline after two decades of cycle abuse — the three-player HBM oligopoly (SK Hynix/Micron/Samsung) now behaves like an oligopoly. Named flow: 16 Strategic Customer Agreements worth ~$100B of contracted multi-year HBM, 2026 HBM supply fully pre-sold under price/volume agreement, HBM pricing +246

  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 and dragging blended margin below the mid-50s.

  Catalyst

  FQ4 earnings (late Sept) — HBM4 qualification and the FY27 sold-out commentary on the SCA book.

  Scenarios · 12mo targets

  $1,100.00 base

  $1,400.00 bull — HBM4 volume ramp lands at 22-23% share, SCA book extends into 2028, and DRAM stays capacity-disciplined — memory re-rates as an infrastructure annuity, not a cycle.

  $650.00 bear — Late-2026 capacity additions plus CXMT commodity DRAM break pricing; a hyperscaler capex digestion quarter turns the sold-out narrative into inventory.

• TSMTaiwan Semiconductor Manufacturingcompute

  hold5.71%86
  flow acceleratingconf

  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 and ~90% advanced-node share. Named flow: 2026 capex guided up to $60-64B; CoWoS booked out through end-2026 at 75-80 KWPM scaling to 115-140 KWPM, with 240-270k wafers/yr outsourced to Amkor/SPIL because demand e

  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; or foundry capex guided down as hyperscaler orders are cut.

  Catalyst

  Q3 results and the October capex update — 2027 capex framing and N2 ramp commentary.

  Scenarios · 12mo targets

  $500.00 base

  $620.00 bull — N2 ramps on schedule with pricing power intact and CoWoS constraint clears; TSMC captures the full 2027 AI-capex step-up and approaches the $4T market-cap case.

  $330.00 bear — A Taiwan geopolitical scare or an AI digestion quarter compresses the multiple even with earnings intact; Arizona cost drag bites gross margin.

• NVDANVIDIAcompute

  hold3.53%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 of 2026 hyperscaler capex anchored on Blackwell/Rubin (MSFT, AMZN, GOOGL, META capex lines all raised again), the Stargate JV, and networking revenue at $14.8B (+199% YoY) proving the platform extends past the GPU. Laffon

  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; or an export-control action removing a material share of demand.

  Catalyst

  Q3 FY27 earnings — Rubin tape-out/shipment timing and data-center growth rate.

  Scenarios · 12mo targets

  $260.00 base

  $330.00 bull — Rubin ships into an undiminished capex line and inference demand (Anthropic at a $65B run rate, +600% YoY) keeps GPU allocation scarce through 2027.

  $155.00 bear — Custom ASIC share crosses the 25% line as TPU/MTIA/Trainium scale, and a single hyperscaler capex trim re-rates the whole complex.

• ASMLASML Holdingcompute

  hold2.90%61
  flow steadyconf

  Funder

  EUV / High-NA monopoly — the single most irreplaceable tool in advanced semis, no commercial-scale alternative at 2nm, multi-year backlog, ~50% gross margin, and an installed-base service annuity that survives order-cycle troughs. Named flow: 60 low-NA EUV units planned for 2026 (+25% units) against $36-40B guided revenue, High-NA moving into high-volume manufacturing, and TSMC's $60-64B capex plu

  Unwind

  Book-to-bill <0.7 for three consecutive quarters; a Dutch/US export-control tightening removing a material share of the orderbook; or a credible non-EUV patterning path at 2nm.

  Catalyst

  Q3 bookings print in October — the single number that either confirms or breaks the 2026 60-unit plan.

  Scenarios · 12mo targets

  $2,000.00 base

  $2,500.00 bull — Bookings re-accelerate on the 2026 EUV unit step-up and High-NA orders convert; the litho monopoly re-rates back toward its 2025 multiple.

  $1,300.00 bear — A China export-control tightening plus a lumpy bookings quarter drops book-to-bill below 0.7 and the score-61 de-rating becomes fundamental.

• AVGOBroadcomcompute

  hold2.86%63
  flow steadyconf

  Funder

  Custom-silicon ASIC leader (~70% share) plus the VMware infrastructure-software lock-in — 69.5% gross margin, huge FCF, $7B+ annual buyback, serial dividend grower, and one of the best capital-allocation records in semis under Tan. Named flow: Alphabet multi-generation TPU, Meta MTIA, and the Anthropic 3.5GW ASIC program; AI segment +143% to $10.8B with $30B+ of bookings. Druckenmiller's full Q2 e

  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; or leverage from the financing program impairing the buyback.

  Catalyst

  Q3 earnings in early September — AI revenue run-rate and the next tranche of ASIC bookings.

  Scenarios · 12mo targets

  $430.00 base

  $530.00 bull — A fourth named hyperscaler ASIC program lands and AI revenue clears a $30B run-rate; VMware renewals hold pricing and the buyback compounds a de-rated share count.

  $265.00 bear — The BofA financing-ceiling narrative becomes a real balance-sheet constraint while a hyperscaler pulls an ASIC program in-house; RSI-36 de-rating extends into an earnings miss.

• ANETArista Networkscompute

  hold2.55%65
  flow steadyconf

  Funder

  Durable hyperscaler-networking franchise and a consistent share-gainer since 2004 — net cash, no debt, and the single-image EOS software moat that keeps switching costs high; Ullal's capital discipline is a two-decade record. Named flow: Microsoft and Meta as named anchor customers funding the first $3B quarter (Q2 revenue $3.04B, +37.7% YoY) and an FY guide raised to $12.6B (+40%); 800G/1.6T camp

  Unwind

  Microsoft+Meta combined concentration >50% with revenue decelerating; Spectrum-X or whitebox displacing EOS in a NAMED hyperscaler footprint; or gross margin breaking below the low-60s on price competition.

  Catalyst

  Q3 earnings — whether any named hyperscaler footprint shifts to Spectrum-X, the one thing that would make Nvidia's $14.8B networking line a displacement rather than a market expansion.

  Scenarios · 12mo targets

  $215.00 base

  $270.00 bull — AI back-end Ethernet wins broaden beyond the two anchors and the $12.6B guide is beaten; EOS proves sticky against Spectrum-X in a competitive bake-off.

  $130.00 bear — Nvidia bundles networking into GPU allocations and takes a named hyperscaler footprint; a ~48x forward multiple on decelerating growth compresses hard.

• GEVGE Vernovaenergy

  hold6.32%61
  flow acceleratingconf

  Funder

  Gas-turbine and grid franchise with genuine pricing power — record ~$176B backlog running through 2031, heavy-duty turbines effectively sold out through 2030, 10-20% price realization on new orders, and a post-spin management team that has already fixed wind losses and turned FCF positive. Named flow: Norway GPFG added post-spin; Morgan Stanley's 38GW US data-center power gap; DOE's ~100GW-by-2030

  Unwind

  Gas-turbine backlog cancellation or a hyperscaler power-capex cut; wind-segment losses widening enough to consume gas-segment profit; or EV/EBITDA breaching the invalidation band with orders flat.

  Catalyst

  Q3 order intake and the SPARK/GRIP award announcements — new gas and grid orders are the backlog confirmation.

  Scenarios · 12mo targets

  $1,150.00 base

  $1,450.00 bull — The 38GW power gap converts into multi-year turbine and grid orders; backlog crosses $200B and the wind segment turns profit-neutral, justifying the 34x.

  $700.00 bear — A hyperscaler power-capex pause stalls order intake at 34x 2026 earnings after a 674% post-spin run — pure multiple compression with no earnings break needed.

• SCCOSouthern Copperenergy

  hold10.66%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 the cycle. Named flow: ~320k-tonne 2026 copper supply deficit with prices near record highs, AI/data-center demand at ~50k tonnes per hyperscale facility and ~30% of new copper demand

  Unwind

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

  Catalyst

  Tia Maria / Los Chancas permitting milestones and the Q3 dividend declaration.

  Scenarios · 12mo targets

  $225.00 base

  $275.00 bull — The structural deficit widens as AI/electrification demand compounds and Tia Maria permits clear — volume growth on top of record realized prices.

  $150.00 bear — A China demand air-pocket cracks copper, or a Peruvian royalty/permitting reversal strands the growth pipeline and forces a dividend cut.

• NEMNewmont Corporationenergy

  hold7.03%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: sustained central-bank gold accumulation as an FX-reserve diversification program, plus the Treasury's announced doubling of long-dated bond buybacks pushing long yields down and the safe-haven bid

  Unwind

  Cost overruns or mine disappointments breaking FCF and the dividend; durable disinflation plus a real-rate surge removing the stagflation-hedge rationale; or a value-destroying acquisition reversing the divestiture discipline.

  Catalyst

  Jackson Hole (Aug 27-29) — Warsh's reaction function is the swing variable for real yields and therefore the gold bid.

  Scenarios · 12mo targets

  $150.00 base

  $190.00 bull — Fed stays pinned into a cracking labor market, real yields fall, and central-bank buying persists — gold breaks higher while AISC discipline holds margins.

  $95.00 bear — Core CPI prints below 2.2% twice, real yields surge on a hawkish Warsh, and the RSI-74 overbought setup unwinds; cost inflation eats the FCF.

• KTOSKratos Defensedefense

  hold11.41%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 the primes cannot easily displace. Named flow: DoD MACH-TB 2.0 $1.45B IDIQ prime ceiling, Project Helios $68.3M single-award hypersonic materials center, USMC Valkyrie MUX mods, and the FY27 DAWG request of $54.6B against a $1.5T FY202

  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 appropriation that strands the DAWG line.

  Catalyst

  NDAA markup / FY2027 appropriations language on unmanned systems, and Q3 earnings for the first margin inflection through the drone ramp.

  Scenarios · 12mo targets

  $70.00 base

  $95.00 bull — DAWG money converts into Valkyrie production orders and drone-segment margin finally inflects with FCF turning positive — the scale story the multiple already assumes.

  $38.00 bear — A CR through the midterms freezes new starts, FCF stays negative, and a 95x forward multiple on sub-scale margins de-rates all the way to asset value.

• HEIHEICOdefense

  hold12.00%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%+ through a disciplined serial-acquisition model (~$1B/yr of bolt-ons at sane prices) with PMA-part economics the OEMs structurally cannot undercut. Named flow: Q2 FY2026 beat with FSG $929M vs $864M consensus and ETG $460M vs $396M, ETG operating in

  Unwind

  Organic growth decelerating below 6% WITH margin compression; a value-destroying large acquisition breaking the serial-acquirer discipline; or forward P/E sustained above 50x while growth decelerates — note the invalidation requires BOTH, and growth is currently accelerating.

  Catalyst

  Q3 FY2026 results — organic growth rate and the pace of bolt-on deployment.

  Scenarios · 12mo targets

  $410.00 base

  $490.00 bull — Aftermarket demand and the ETG ramp keep organic growth in the teens while the acquisition machine deploys $1B+ at disciplined prices — the multiple is earned.

  $265.00 bear — A commercial-aero cycle turn drops organic growth below 6% while the 54x forward multiple is still in place — both invalidation legs fire together.

• LLYEli Lillybiology

  hold3.46%74
  flow acceleratingconf

  Funder

  Premier pharma compounder — the Mounjaro/Zepbound tirzepatide franchise (~60% US share; Q2 Mounjaro $9.9B, Zepbound $4.9B) on 85.8% gross margin, plus a manufacturing moat from the $27B capacity build that rivals cannot replicate quickly. Named flow: Foundayo (orforglipron) took Europe's FIRST oral GLP-1 approval on Aug 14 in the UK, with no food restrictions versus Novo's Wegovy, and retatrutide

  Unwind

  Tirzepatide US script share falling >5pts to Novo over two quarters; retatrutide Phase 3 disappointing on safety or durability; or IRA/MFP negotiation reaching tirzepatide with a price cut that breaks the margin structure.

  Catalyst

  Retatrutide Phase 3 readouts and the EU/US orforglipron launch trajectory; CMS negotiation list updates.

  Scenarios · 12mo targets

  $1,500.00 base

  $1,850.00 bull — Retatrutide confirms best-in-class efficacy and orforglipron scales the oral TAM globally — Lilly owns both the injectable and oral legs of a $92-204B obesity market.

  $950.00 bear — CagriSema or an oral rival takes measurable US script share while IRA/MFP negotiation reaches tirzepatide — volume growth meets price compression at once.

• VRTXVertex Pharmaceuticalsbiology

  hold2.88%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 — a franchise insulated from the rate vise because it never needs the capital markets. Named flow: the $10B Crinetics acquisition (Jul 30) deploys that cash into endocrinology, plus Casgevy commercializi

  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 with growth stalled; or a suzetrigine reimbursement failure.

  Catalyst

  Suzetrigine launch metrics and Casgevy uptake figures at Q3; Crinetics integration milestones.

  Scenarios · 12mo targets

  $630.00 base

  $760.00 bull — Suzetrigine scales into the non-opioid pain gap and Casgevy reimbursement broadens — the non-CF leg finally re-rates the whole company off a CF-monopoly multiple.

  $430.00 bear — Suzetrigine launch disappoints on payer coverage and Crinetics proves a $10B overpay, leaving a rate-pressured single-franchise stock.

• ARGXargenxbiology

  hold2.82%70
  flow acceleratingconf

  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, so it does not depend on the capital markets in a 2.41% real-yield world. Named flow: FDA expanded gMG approval to all serotypes including seronegative (an 18% TAM expansion), the MG market is projected to g

  Unwind

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

  Catalyst

  Forte tender close (~Aug 26) and the next Vyvgart quarterly revenue print with subcutaneous conversion rates.

  Scenarios · 12mo targets

  $1,250.00 base

  $1,500.00 bull — Subcutaneous Hytrulo conversion accelerates and CIDP share compounds with Sanofi out of the way; myositis/Sjogren's readouts extend the label runway.

  $780.00 bear — A pipeline indication miss plus an FcRn entrant taking share stalls Vyvgart growth while an RSI-73 stretched multiple unwinds.

Warnings

• stale_policy guard: MKL open → hold (delta zeroed)
• stale_policy guard: V open → hold (delta zeroed)
• §2 cap re-applied after pillar tilt: HEI
• §5.4 dry-powder: cash_pct 23% above base band top 15% allowed by §3b unfilled-sleeve headroom (+17.5pp, eff top 32.5%)
• cash_band: final cash 22.09% above neutral base band top 15% (escalator active: extension; §3b dry-powder headroom +17.5pp (unfilled opportunistic sleeve → cash)) (§5.4)
• position_count: 14 < 16 after mechanical closes (§2 floor) — book under-populated; strategist must open replacements next run
• live_sizing: 1 ticket(s) re-expressed against the executed ledger; dropped add SCCO (live already at target) (§6.6)
• live_sizing: dropped add SCCO (Δ2.36pp vs intent) — live weight 8.91% already satisfies target 10.66% (§6.6; the ledger, not intent, is what a trade moves)
• open_universe: MKL score=51 < 60 and no §6.3a basis — dropped
• buy guard dropped 2 open(s): MKL, V (§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.