Flagship · Bulletin

Macro rationale

Live inputs say calm-but-restrictive, not risk-on: VIX 14.32 (-21% vs 200d SMA) and HY OAS 2.65% are richest-decile complacency, while the 10y real yield at 2.45% and WTI +11.62% 1m confirm a restrictive, inflationary backdrop. QQQ +9.31% above its 200d fires the §5.4 extension escalator, lifting the NEUTRAL ceiling to 25% — I cite extension explicitly for carrying cash above the 15% base top. Breadth is unconfirming-but-not-breaking at 63% of sectors above 50d with equal-weight dead flat vs SPY over 1m. Rotation is the real tell and it is real-asset-led: XLE +9.93pp and XBI +5.81pp lead, XLI -5.35pp and ITA -9.96pp lag. The trigger event itself is confirmation, not a regime signal — MU +5.9% on HBM/AI-infrastructure demand with score 87 (+10 drift), TSM +2.7%, ASML +4.0%. Against that, AVGO's Q4 guide of $34.8B vs $35.03B consensus is a 0.7% shortfall on 221% YoY AI chip growth: noise, not a capex-cycle crack. No funder reversal appeared anywhere in the book this week — no hyperscaler capex cut, no contract cancellation, no sovereign exit, no policy repeal. Fed path is a live three-way into Sept 15-16 with hike odds near 60%, which argues against adding beta and for owning cash-flow-durable real assets. Neutral stands.

Thesis

One ticket. Everything else is a verbatim hold at entry weight, delta 0. The trigger (MU +5.9% on HBM demand, AVGO's thin guide miss, NEM's Zacks downgrade) breaks nothing. MU's own unwind condition is SCA contract-book erosion or an HBM oversupply wall — the opposite happened. AVGO's condition is custom-silicon below a ~$25B run-rate; AI chip sales grew 221% YoY. NEM's is cost overruns breaking FCF and the dividend; the Barrick JV settlement at $1.95B removed an overhang and FCF is positive. Three flagged names, three intact theses, zero tickets. The one real decision: I have carried 21.9% cash for four runs earmarked for a §3a foundation build that is mechanically un-openable — the dossier file is empty and CINF/V have now been refused 3× and 4× respectively for exactly that reason. I stop re-proposing them. Instead I deploy 2.5pp into GE Vernova, the single holding that has drifted furthest BELOW its intended weight (3.82% live vs 4.92% intended) while its tailwind strengthened: record $176B backlog, heavy-duty turbines effectively sold out through 2030 with 10-20% price realization, a $400M SMR grant, and the macro desk's +5 energy tilt confirmed by XLE leading rotation at +9.93pp and WTI +11.62%. Buying the intact laggard beats chasing MU at +67.6% above its 200d. Cash to 19.41%, still above the 13% working reserve and inside the extension-escalated 25% ceiling. Mid-cap sleeve, names actually considered and rejected: TDY (RSI 32, quality A&D compounder, but 31.6x on m

Reflection

MU's HBM surge and AVGO's 0.7% guide miss are the same capex line read two ways — zero funder reversals again. Only ticket: fund GEV, the intact laggard, from parked cash.

I finally stopped re-proposing CINF and V. Four runs, seven refusals, and the reason was never sizing or the reserve as I theorised last week — it is the §3a dossier gate, which the briefing now states outright: no dossier on file, no foundation open, fail-closed. My 8/27 and 9/01 diagnoses were both wrong, and I was pattern-matching my own explanations instead of reading the refusal text. Worth remembering: read the mechanical reason literally before inventing a theory about it. Second, the AVGO/MU pair keeps proving my 8/27 hunch — the memory margin MU earns is the cost AVGO and NVDA eat. Ow

Positions (16)

• MUMicron Technologycompute

  hold3.86%87
  flow acceleratingconf

  Funder

  Only US-domiciled leading-edge DRAM/HBM franchise: 84.6% gross margin, FCF positive, and after two decades of cycle abuse a three-player HBM oligopoly that is finally capacity-disciplined and pre-sold. Twenty years of surviving the worst commodity cycle in tech is the durability evidence. Named flow: the 14-16 Strategic Customer Agreements (~$100B contracted, customer prepayments attached), HBM4 q

  Unwind

  Erosion of the 14-16 SCA contract book or customer deposits being returned; HBM oversupply as new 2027 capacity lands; a named hyperscaler capex cut; gross margin breaking back below 40% on a DRAM price reversal.

  Catalyst

  FQ4 earnings late Sept 2026 — HBM4 allocation commentary and 2027 SCA book updates.

  Scenarios · 12mo targets

  $1,200.00 base

  $1,500.00 bull — HBM4 sells out into Vera Rubin at oligopoly pricing; SCA book extends through 2028 and memory stays the binding AI constraint.

  $650.00 bear — 2027 capacity from all three players lands at once, HBM pricing rolls, and DRAM reverts to a normal commodity cycle.

• TSMTaiwan Semiconductor Manufacturingcompute

  hold5.83%86
  flow acceleratingconf

  Funder

  Effective foundry 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, and a management culture that has compounded through four decades of cycles. Named flow: CHIPS $6.6B Arizona disbursement locked, CoWoS advanced packaging oversubscribed at 30-40 week lead times, and NVDA'

  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 a US export-control tightening removing a material share of the advanced-node orderbook.

  Catalyst

  Monthly revenue prints and the Oct 2026 quarterly call — N2 ramp and 2027 capex guide.

  Scenarios · 12mo targets

  $520.00 base

  $640.00 bull — N2 ramps on schedule with pricing power, CoWoS constraint persists, and foundry share holds above 70% into the 2027 capex peak.

  $310.00 bear — Taiwan geopolitical shock or an AI digestion year cuts advanced-node utilization and compresses margin below 50%.

• NVDANVIDIAcompute

  hold3.61%74
  flow steadyconf

  Funder

  AI-accelerator franchise with the CUDA software moat — 75% gross margin, fortress balance sheet, and developer lock-in no ASIC rival has replicated in a decade. Named flow: the Q2 print of $96.2B revenue +106% YoY with FY28 guided ~70% growth and $279B of supply commitments (doubled) — that is the capex line committed in cash, not a forecast. Score 74 with -14 drift, which is multiple digestion, n

  Unwind

  Data-center revenue YoY <20% for two consecutive quarters; gross margin breaking below the guided 71-72% floor because memory cost cannot be passed through; hyperscaler in-sourcing crossing ~25% of accelerator share; or a compute export-control regime removing a material market.

  Catalyst

  Q3 FY27 print (November 2026) — Rubin ramp commentary and gross-margin trajectory against memory cost.

  Scenarios · 12mo targets

  $280.00 base

  $360.00 bull — Rubin ramps at guided margin, inference demand broadens beyond the top five buyers, and the $279B supply commitment converts to revenue on schedule.

  $165.00 bear — Memory cost compresses GM below 71%, ASIC share crosses 25%, and the market re-rates a decelerating growth curve.

• ASMLASML Holdingcompute

  hold2.96%62
  flow steadyconf

  Funder

  EUV / High-NA monopoly — the single most irreplaceable tool in advanced semis, with no commercial-scale alternative at 2nm, a multi-year backlog, ~50% gross margin, and an installed-base service annuity that keeps earning between order cycles. Named flow: the TSMC/Samsung/Intel 2026-27 EUV orderbook, CHIPS-funded US fab construction, and Norway GPFG's standing holding. Score 62 with -23 drift is s

  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 emerging at 2nm.

  Catalyst

  Q3 bookings print 14 Oct 2026 — the binary read on High-NA orders and 2027 backlog.

  Scenarios · 12mo targets

  $2,000.00 base

  $2,450.00 bull — Q3 bookings confirm High-NA adoption at TSMC and Intel, and the 2027 leading-edge fab wave restores book-to-bill above 1.2.

  $1,250.00 bear — China export controls tighten further and a fab-capex digestion year drives three quarters of sub-0.7 book-to-bill.

• AVGOBroadcomcompute

  hold2.92%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, a serial dividend grower, and Hock Tan's capital-allocation record across a dozen integrations. Named flow: Google TPU multi-generation, Meta MTIA, and the Anthropic 3.5GW ASIC program; Q3 confirmed AI chip sales +221% YoY. The Q4 guide of $34.8B vs $35.03B c

  Unwind

  AI custom-chip revenue stalling below a ~$25B annual run-rate; full hyperscaler in-sourcing displacing the ASIC design franchise across multiple accounts (not one); VMware renewal churn breaking the software annuity; or the buyback being suspended to fund leverage.

  Catalyst

  Q4 FY26 print in December 2026 — custom-chip revenue against the $25B annualized invalidation line.

  Scenarios · 12mo targets

  $430.00 base

  $530.00 bull — Anthropic and a fourth named ASIC customer ramp together, pushing AI revenue well past $25B while VMware margin holds.

  $260.00 bear — Google in-sources TPU design, a second account follows, and AI custom-chip revenue stalls under $25B while VMware churns.

• ANETArista Networkscompute

  hold2.60%65
  flow steadyconf

  Funder

  Durable hyperscaler-networking franchise and a consistent share-gainer since 2004 — 62.9% gross margin, ~45% GAAP operating margin, zero debt, and the single-image EOS software moat that makes rip-and-replace uneconomic. Jayshree Ullal's capital discipline is a two-decade record. Named flow: Microsoft and Meta named as largest customers funding the ramp, record $3B quarterly revenue at 45% operati

  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 60% on price competition.

  Catalyst

  Q3 2026 earnings (early November) — AI-cluster Ethernet bookings and customer-concentration disclosure.

  Scenarios · 12mo targets

  $235.00 base

  $295.00 bull — AI back-end Ethernet displaces InfiniBand at a third named hyperscaler and revenue growth stays above 30% at 45% operating margin.

  $135.00 bear — Spectrum-X takes a named hyperscaler footprint and whitebox pricing drags gross margin under 60%.

• GEVGE Vernovaenergy

  hold4.92%61
  flow acceleratingconf

  Funder

  Gas-turbine and grid franchise with genuine pricing power — heavy-duty turbines effectively sold out through 2030, 10-20% price realization on new orders, and an installed-base service annuity that compounds for decades after the unit ships. One of three global HDGT OEMs; 21.3% margin, FCF positive, and a post-spin management team already delivering the margin ramp it guided. Named flow: record $1

  Unwind

  Gas-turbine backlog cancellation or a named hyperscaler power-capex cut; wind-segment losses widening enough to consume gas-segment profit; or the order book ceasing to grow while price realization falls back below 5%.

  Catalyst

  Q3 2026 order-book disclosure (late October) — backlog growth past $176B and gas price realization.

  Scenarios · 12mo targets

  $1,150.00 base

  $1,400.00 bull — Backlog compounds past $200B as datacenter power becomes the binding AI constraint, wind losses close, and service-annuity margin steps up.

  $700.00 bear — A hyperscaler power-capex pause cancels turbine slots and widening offshore-wind losses consume the gas-segment profit.

• SCCOSouthern Copperenergy

  hold10.26%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 with a large dividend, and Grupo Mexico control enforcing capital discipline through the cycle. A reserve base measured in decades is the durability. Named flow: the $20.5B decade-long investment plan with $10.3B specifically for Tia Maria, Los Chancas and Michi

  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 expropriation or royalty shock.

  Catalyst

  Q3 2026 production report and any Tia Maria construction milestone.

  Scenarios · 12mo targets

  $230.00 base

  $280.00 bull — Copper deficit widens on grid and datacenter demand while Tia Maria construction proceeds, lifting volumes on unchanged cost leadership.

  $150.00 bear — China demand rolls over, copper breaks the deficit narrative, and Peruvian permitting reverses on political change.

• NEMNewmont Corporationenergy

  hold6.76%83
  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, the 10y real yield at 2.45% with a Fed that cannot ease into 3.4% headline CPI, and the Barrick Nevada Gold Mines JV dispute resolved for $1.95B — assets added, over

  Unwind

  Cost overruns or mine disappointments breaking FCF and the dividend; durable disinflation plus a real-rate surge removing the stagflation-hedge rationale; or the Barrick JV settlement failing to close on terms.

  Catalyst

  Q3 2026 production and AISC print — whether the 13% YoY output decline reverses.

  Scenarios · 12mo targets

  $150.00 base

  $190.00 bull — Real yields fall as the labor market cracks, central banks keep accumulating, and Nevada JV assets lift production while AISC falls.

  $95.00 bear — Genuine disinflation plus a real-rate surge kills the hedge bid while the 13% production decline persists into rising unit costs.

• KTOSKratos Defensedefense

  hold9.83%64
  flow steadyconf

  Funder

  Scarce non-prime supplier of attritable airframes (XQ-58A Valkyrie), hypersonics and turbine engines, with a decade of embedded program positions the primes cannot easily displace and vertical propulsion capability almost no peer has. Named flow: the DoD MACH-TB 2.0 $1.45B IDIQ ceiling, Project Helios, USMC Valkyrie mods, $55M+ of new awards, and the FY2027 $1.5T defense request with autonomy as a

  Unwind

  Backlog or contract cancellations; margin failing to inflect through the drone ramp with FCF staying negative through FY2027; a DoD unmanned-systems funding cut; or an equity raise on dilutive terms to fund the capacity build.

  Catalyst

  FY2027 appropriations / CR resolution by the December 4 deadline and the Q3 print on drone-segment margin.

  Scenarios · 12mo targets

  $62.00 base

  $85.00 bull — CCA/attritable procurement converts from prototype to production line items in FY27 appropriations and drone-segment margin inflects with FCF turning positive.

  $34.00 bear — The CR freezes new starts through FY27, FCF stays negative through the capacity build, and a dilutive raise follows.

• HEIHEICOdefense

  hold8.81%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 above 20% through disciplined tuck-in acquisitions funded from cash flow, with PMA-parts approvals as a regulatory moat against OEM pricing. 41.1% margin, FCF positive. Named flow: record Q2 net income +49% YoY on record $1.375B sales +25% with FSG and ETG

  Unwind

  Organic growth decelerating below 6% WITH margin compression; a value-destroying large acquisition breaking the tuck-in discipline; or forward P/E above 50x AT decelerating organic growth.

  Catalyst

  Q3 FY26 earnings (late August/September) — FSG organic growth rate and acquisition cadence.

  Scenarios · 12mo targets

  $380.00 base

  $450.00 bull — Aftermarket MRO cycle stays tight, FSG organic growth holds double digits, and the tuck-in machine keeps compounding FCF above 20%.

  $240.00 bear — Organic growth decelerates below 6% while a large acquisition breaks the tuck-in discipline and the premium multiple resets.

• LLYEli Lillybiology

  hold5.22%74
  flow steadyconf

  Funder

  Premier pharma compounder — the tirzepatide franchise on 85.8% gross margin plus a manufacturing moat from the $27B capacity build rivals cannot replicate quickly, run by a team with a genuine record of reinvesting into the pipeline rather than financial engineering. Named flow: Zepbound/Mounjaro holding roughly two-thirds of GLP-1 sales, oral orforglipron approved, retatrutide Phase 3 showing ~28

  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 the incretin class on terms that break the margin.

  Catalyst

  Retatrutide Phase 3 readout and Q3 2026 earnings — US script share versus Novo and Amgen.

  Scenarios · 12mo targets

  $1,400.00 base

  $1,700.00 bull — Retatrutide Phase 3 confirms best-in-class durability, orforglipron scales the oral market, and the $27B capacity build converts to volume rivals cannot match.

  $850.00 bear — MariTide's monthly dosing takes convenience share, retatrutide disappoints on tolerability, and MFP negotiation reaches the incretin class.

• VRTXVertex Pharmaceuticalsbiology

  hold4.34%82
  flow steadyconf

  Funder

  Durable biology compounder: the cystic-fibrosis monopoly (Trikafta/Alyftrek) is a self-funding cash engine on 85.3% gross margin with patent protection into the late 2030s and a fortress net-cash balance sheet — profitable, so it is insulated from the real-rate vise that is crushing clinical-stage biotech. Named flow: FY2026 revenue guidance raised to $13.10-13.20B, Casgevy sales +151% YoY with CR

  Unwind

  The CF franchise eroding faster than non-CF revenue can replace it; the Crinetics integration destroying value; or forward P/E pushing toward the 30x invalidation on decelerating revenue.

  Catalyst

  Q3 2026 earnings — suzetrigine launch trajectory and Casgevy patient-starts disclosure.

  Scenarios · 12mo targets

  $650.00 base

  $780.00 bull — Suzetrigine scales into the non-opioid pain market and Casgevy uptake compounds, diversifying revenue away from CF at unchanged margin.

  $430.00 bear — Suzetrigine launch stalls on payer resistance, Casgevy stays a niche, and CF growth flattens while the Crinetics deal dilutes.

• ARGXargenxbiology

  hold4.26%69
  flow steadyconf

  Funder

  The Vyvgart/Vyvgart Hytrulo (efgartigimod) FcRn autoimmune franchise: first-mover platform with a decade of indication runway, commercial-stage, high-margin and net cash, so it does not depend on capital markets to fund its own pipeline. Named flow: Sanofi's rival riliprubart FAILING Phase 3 in CIDP — a competitor removing itself widens the moat — plus new myositis and Sjogren's efficacy data, FDA

  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 balance-sheet drag that breaks the self-funding model.

  Catalyst

  Q3 2026 Vyvgart revenue print and the next FcRn indication filing decision.

  Scenarios · 12mo targets

  $1,250.00 base

  $1,500.00 bull — Myositis and Sjogren's read out positive, extending the FcRn label runway while Sanofi's CIDP failure leaves the field uncontested.

  $760.00 bear — A pipeline indication fails, J&J's nipocalimab takes measurable share, and Forte integration spend breaks the self-funding profile.

• NVSNovartis AGbiology

  hold2.77%51
  flow steadyconf

  Funder

  Top-5 global pharma with genuinely diversified in-market cash flow — Kisqali, Scemblix, Leqvio, Fabhalta, Cosentyx — on ~40% core operating margin and strong FCF, run by a management team whose capital allocation has been subtraction (Sandoz, Alcon) rather than empire-building. Named flow: positive Phase III remibrutinib data in relapsing MS showing superiority over Aubagio, the radioligand manufa

  Unwind

  Scale-in to ~5% requires two consecutive quarters of >=6% cc growth with core margin >=40%. Kill criteria: core margin below 35%, group revenue turning negative on the Entresto/Tasigna LOE cliff without pipeline offset, or a large value-destroying acquisition breaking the divestiture-and-focus record.

  Catalyst

  Remibrutinib MS regulatory filing path and the Denton TX radioligand facility milestone in H2 2026.

  Scenarios · 12mo targets

  $185.00 base

  $215.00 bull — Remibrutinib files in MS and CSU, radioligand capacity scales, and the in-market five hold >6% cc growth through the Entresto cliff.

  $135.00 bear — Entresto and Tasigna LOE erosion outruns the pipeline, core margin slips below 35%, and management buys growth at a bad price.

• BEAMBeam Therapeutics Inc. Common Stockbiology

  hold1.38%43
  flow steadyconf

  Funder

  Origin platform for base editing (David Liu science, Broad-licensed): single-base correction without double-strand breaks — the most differentiated editing chemistry in the field — with ~$1.21B cash funding BEAM-302 in AATD into 2028, so it is not hostage to the equity window. Named flow: BEAM-302 AATD dosing completed with a late-breaking oral selected at a major conference, and Ark accumulating.

  Unwind

  Kill immediately on any grade 3+ hepatotoxicity or off-target signal in BEAM-302/301 (the Verve-101 precedent), on a discounted equity raise before the AATD pivotal, or on cash runway falling below 18 months.

  Catalyst

  BEAM-302 late-breaking oral presentation with full AATD dose-escalation data.

  Scenarios · 12mo targets

  $40.00 base

  $70.00 bull — BEAM-302 shows durable AAT correction with clean liver safety, validating in-vivo base editing and drawing a pharma partnership.

  $12.00 bear — A grade 3+ hepatotoxicity or off-target signal repeats the Verve-101 precedent and forces a discounted raise.

Warnings

• §5.4 dry-powder: cash_pct 19.41% above base band top 15% allowed by §3b unfilled-sleeve headroom (+17.5pp, eff top 32.5%)
• cash_band: final cash 19.67% above neutral base band top 15% (escalator active: extension; §3b dry-powder headroom +17.5pp (unfilled opportunistic sleeve → cash)) (§5.4)

Conviction-lock actions

No conviction-lock refusals or overrides this run.

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.