Flagship · Bulletin

Macro rationale

Event-driven patch run, scope locked to [ASML,MU,TSM,NVDA,AVGO,ANET,FN]; all other names held unchanged per the ad-hoc trigger. Regime stays NEUTRAL with a defensive lean: credit tight (HY OAS 2.71 below 90d sma), VIX only +4.5% vs 200d, but the extension escalator is ACTIVE (QQQ +12.65% >200d) so cash holds the upper band, not the floor. The Warsh Fed delivered a hawkish hold (dot median 3.8%, CME Sept hike odds 72.8%) which keeps a discount-rate headwind on long-duration compute multiples — exactly the force driving the ASML/NVDA/AVGO score drift this week, NOT a demand/flow break. The trigger itself — MU's blowout Q3 ($41.46B vs $36B est), $50B Q4 guide, and 16 multi-year Strategic Customer Agreements worth ~$100B — is the cleanest single confirmation this cycle that the AI-memory/silicon capex flow is structural, not episodic. That read-through is bullish for the whole physical-AI stack I hold (TSM foundry, ASML litho for memory-fab expansion, FN/ANET cluster interconnect), so the right response to the event is to lean into the confirmed name (MU) and hold — not trim — the franchises the rate move is compressing.

Thesis

Event-driven patch, scope locked to [ASML,MU,TSM,NVDA,AVGO,ANET,FN]; the other 10 names are held unchanged at delta 0 per the ad-hoc trigger, briefings spliced from the prior Sunday deep. The trigger is MU's blowout: Q3 $41.46B vs $36B est (verified +17.5% to ~$1,232 on Polygon), $50B Q4 guide, and 16 multi-year Strategic Customer Agreements worth ~$100B in contracted revenue. That is the single cleanest accelerating AI flow this cycle and it is structural, not episodic — so I made exactly one move: add +1.5pp to MU (3.76%->5.26%), leaning into the confirmed name. Adds are not lock-restricted, so the 1d unlock=none does not block it. The read-through from MU's $100B demand + capacity build is positive for the rest of the physical-AI stack I hold (TSM foundry volume, ASML memory-fab litho orders, FN/ANET cluster interconnect), which is why I am holding — not trimming — ASML/NVDA/AVGO despite their score drift: that drift is the Warsh-Fed discount-rate compression on long-duration multiples, not a demand or named-funder break (EUV monopoly, $725B capex, 70% ASIC share all intact). Cash sits at 6.62%, inside the neutral 5-15% band. Book remains 17 names — the patch scope forbids opening an 18th outside the named list (all 7 already held), so the count-band gap carries to the next full Sunday deep.

Reflection

MU's $100B contracted-revenue blowout is the cleanest structural AI-memory flow this cycle; compute score drift elsewhere is rate-driven multiple compression, not a demand break.

The MU print resolves a tension I've carried for weeks: was the June compute score drift (ASML -22, NVDA/AVGO -16) a flow reversal or a discount-rate move? MU's $41.46B/$50B-guide/$100B-agreements answers it — underlying AI capex demand is accelerating, so the drift is the Warsh-Fed discount rate on duration, exactly as I hypothesized. The read-through chain (MU demand -> memory-fab capacity -> ASML litho orders) is a reason to HOLD ASML through its score slide rather than trim to monitoring weight as the B1 briefing suggested. Unresolved: book is stuck at 17 vs the 18-20 band and patch scope

Positions (17)

• MUMicron Technologycompute

  hold4.80%77
  flow acceleratingconf

  Funder

  US HBM3e/HBM4 supplier in a capacity-disciplined memory cycle, sold out through 2026-27. Trigger event: blowout Q3 $41.46B vs $36B est, Q4 guide ~$50B vs $43B Street, and 16 multi-year Strategic Customer Agreements worth ~$100B in contracted revenue signed (incl. Anthropic 6/22) — structural AI-memory demand confirmation, underpinned by the ~$725B 2026 hyperscaler capex line. Goldman names memory

  Unwind

  DRAM/HBM ASP roll-over as capacity-discipline breaks, a hyperscaler HBM order cancellation, or the $100B contracted book failing to convert to shipped revenue. Memory is cyclical — a demand air-pocket that pushes the cycle past peak would unwind the thesis; the $1,200 call-wall is near-term technical, not a thesis break.

  Catalyst

  Q4 FY26 print (~Sep 2026) — test whether the ~$50B guide and $100B contracted book convert; HBM4 qualification milestones.

• TSMTaiwan Semiconductor Manufacturingcompute

  hold8.31%83
  flow steadyconf

  Funder

  Leading-edge foundry monopoly (<3nm), >50% gross margin, prodigious FCF, disciplined capex — routes the world's AI silicon for decades. Goldman's $920B-$1.4T 2027 AI-capex forecast names TSMC explicitly; CHIPS $6.6B Arizona disbursement locked; MU's $100B AI-memory agreement is a direct positive read-through for total AI silicon volume. Pillar anchor, score 83 stable (+1).

  Unwind

  A book-to-bill break at leading edge, a hyperscaler in-sourcing wave that bypasses TSMC nodes, or a Taiwan-geopolitical shock severing the supply chain. Apple's Intel-foundry dabble is marginal diversification risk, not a break.

  Catalyst

  TSMC Q2 2026 earnings + capex guide; 3nm price-hike realization 2026-27.

• NVDANVIDIAcompute

  hold3.92%72
  flow steadyconf

  Funder

  CUDA/accelerator software moat, ~75% gross margin, fortress balance sheet. ~$725B 2026 hyperscaler capex anchored on Blackwell/Rubin (MSFT $190B, AMZN $200B, GOOGL $180-190B, META $125-145B); Stargate $500B JV; MU blowout confirms the AI-infra spend NVDA sits at the center of. Score 72 drift is rate-driven multiple compression, not a flow reversal.

  Unwind

  ASIC/custom-silicon displacement materially eroding DC GPU share, a hyperscaler capex cut, or a compute export-control shock (§174/CHIPS amendment). Score drift alone is not an exit.

  Catalyst

  NVDA Q2 FY27 print; Rubin ramp commentary; any new China export-control action.

• ASMLASML Holdingcompute

  hold3.94%63
  flow steadyconf

  Funder

  EUV lithography monopoly — the single most irreplaceable tool in advanced semis, ~50% gross margin, multi-year backlog, no commercial-scale alternative. Norway GPFG holds, Coatue/Third Point added Q1; TSMC/Samsung/SMIC 2025-27 EUV orderbook. MU's $100B demand + capacity expansion is a direct positive read-through for memory-fab litho orders. Score 63 drift is rate-driven, no bookings break.

  Unwind

  A book-to-bill collapse, a sustained export-control tightening that caps China litho revenue, or a memory/foundry capex retrenchment cancelling EUV orders. Hold the monopoly through the discount-rate compression.

  Catalyst

  ASML Q2 2026 bookings (book-to-bill is the key tell); China export-control updates.

• AVGOBroadcomcompute

  hold3.04%62
  flow softeningconf

  Funder

  Custom-silicon ASIC leader (~70% share) + VMware software, huge FCF, $7B+ annual buyback, serial dividend grower. GOOG TPU multi-gen, META MTIA, Anthropic 3.5GW ASIC contracts; Q2 confirmed AI segment +143% to $10.8B and $30B+ bookings. Score 62 drift reflects inference-share narrative favoring NVDA, not a contract loss.

  Unwind

  ASIC custom-chip revenue stalling below the ~$25B run-rate, a hyperscaler fully in-sourcing the design stack, or VMware FCF deterioration. One more leg of score decline near the 60 floor warrants a re-examine, not an auto-trim.

  Catalyst

  AVGO Q3 print — reaffirmation of custom-chip revenue trajectory above the $25B line.

• ANETArista Networkscompute

  hold2.61%67
  flow steadyconf

  Funder

  Durable hyperscaler-networking franchise, share-gainer since 2004, ~65% gross margin, net cash, single-image EOS software moat, Ullal-led capital discipline. Microsoft & Meta named as largest customers funding 35% YoY revenue growth; Morgan Stanley raised PT $180->$190 Overweight (6/12). MU's AI-cluster demand confirms the Ethernet build-out flow.

  Unwind

  A hyperscaler-concentration loss (MSFT/META cutting Arista share), Spectrum-X/proprietary-fabric displacement, or AI-cluster capex deceleration. Lock=15d unlock=none — no trim warranted regardless; the elevated P/E is a watch item, not an invalidation.

  Catalyst

  ANET Q2 print — hyperscaler bookings and 800G Ethernet pull-through.

• FNcompute

  hold1.86%63
  flow softeningconf

  Funder

  Optical transceiver/packaging manufacturer interconnecting hyperscaler AI clusters; capital-light, debt-free, extracts hyperscaler capex without silicon R&D risk. AWS dominant customer drove ~35.9% YoY growth; 800G/1.6T demand from NVDA/AMD GPU clusters. MU's confirmed AI cluster build-out is a positive read-through for optical interconnect volume.

  Unwind

  A >5% sequential Datacom revenue decline confirming demand normalization, AWS in-sourcing optical assembly, or 800G/1.6T transition stalling. Post-beat selloff was sentiment; score 63 near the floor keeps this on a tight watch.

  Catalyst

  Next Datacom sequential print — the >5% decline trigger; 1.6T ramp commentary.

• CCJCamecoenergy

  hold7.51%65
  flow softeningconf

  Funder

  Tier-1 Saskatchewan uranium franchise + Westinghouse 49% JV with Brookfield; commodity tail-risk hedge slot (§5.3). Utility contract book at $60+/lb vs term market ~$93/lb; Kazatomprom 8Mlb output cut sustains structural deficit; 75 reactors under construction; hyperscaler nuclear-PPA rush. Out of patch scope — held unchanged.

  Unwind

  Uranium term-price collapse below contract floors, a major reactor-build cancellation wave, or utility contracting freeze. Score 65 drift is spot softness, not a thesis break.

  Catalyst

  Uranium term-contracting cycle data; CCJ Q2 print.

• GEVGE Vernovaenergy

  hold8.93%63
  flow steadyconf

  Funder

  Gas-turbine + grid franchise with pricing power — ~$163B backlog sold through 2028, 100GW turbine backlog, 10-20% price realization. Norway GPFG added post-spin; hyperscaler grid pull-through; Goldman flags GEV as the AI-capex power bottleneck; potential Iran-reconstruction order catalyst. Out of patch scope — held unchanged.

  Unwind

  Backlog cancellation, gas-turbine margin guidance cut, or wind-segment drag overwhelming grid/turbine economics. Score 63 drift despite supportive news warrants monitoring, not a trim.

  Catalyst

  GEV Q2 print — order-book + electrification backlog disclosure.

• SCCOSouthern Copperenergy

  hold6.95%63
  flow steadyconf

  Funder

  Lowest-cost copper producer, vast reserves, fat margins, large dividend; commodity tail-risk hedge slot (§5.3/10.2). Copper at all-time highs >$6.50/lb (vs $5 entry trigger); AI/data-center ~30% of new copper demand by 2030 (BloombergNEF); BHP copper now outpacing iron-ore profits. Out of patch scope — held unchanged.

  Unwind

  Copper price break below cost-curve support on a China-demand collapse, or Peru permitting/expropriation shock on Tia Maria/Los Chancas. Score 63 stable.

  Catalyst

  Copper price action; Tia Maria/Los Chancas permitting update.

• MPMP Materialsenergy

  hold6.14%
  flow steadyconf

  Funder

  Only scaled US rare-earth + magnet producer, DoD-backstopped moat vs China's 98% magnet control. DoD $400M Series A preferred + $150M loan + 10-yr NdPr price floor + 10-yr magnet offtake for the 10X facility; MOFCOM No.23 (6/22) targeting MP confirms its strategic centrality. Out of patch scope — held unchanged.

  Unwind

  TTM FCF staying negative with magnet-segment ramp stalling, or a DoD funding/price-floor repeal. Score gap this week is a data issue, not a confirmed break.

  Catalyst

  MP Q2 print — TTM FCF + magnet-ramp confirmation against the entry trigger.

• NEMNewmont Corporationenergy

  hold5.02%87
  flow softeningconf

  Funder

  Largest gold producer, tier-1 long-life assets, strong post-Newcrest FCF and dividend, improving cost discipline; tail-risk equity hedge (§5.3/10.2). Central-bank gold accumulation; structural hedge against a frozen-Fed stagflation regime and the out-of-the-money Fed put. C-suite internal promotions signal operational continuity. Out of patch scope — held unchanged.

  Unwind

  Gold breaking durably below bear-market lows on a confirmed disinflation + geopolitical de-escalation combo (the named exit-invalidation scenario), plus NEM cost/FCF deterioration. Lock=20d unlock=none; held through the drawdown by design as the hedge slot.

  Catalyst

  Gold price stabilization; NEM Q2 cost/FCF print.

• KTOSKratos Defensedefense

  hold10.05%
  flow steadyconf

  Funder

  Tactical drones (XQ-58 Valkyrie), hypersonics, turbine engines — DAWG-adjacent defense-tech. DoD MACH-TB 2.0 $1.45B IDIQ ceiling, Project Helios $68.3M, USMC Valkyrie mods; FY27 DAWG ~$54.6B line + 200k-system Drone Dominance program; SpaceX IPO + OpenSpace satellite ground-system growth. Out of patch scope — held unchanged.

  Unwind

  A material drone/hypersonics contract loss, FY27 reconciliation-bill failure cutting the DAWG line, or margin deterioration. The 4.4% Iran-peace-deal dip is sentiment on target-drone demand fears, not a contract change.

  Catalyst

  Q2 earnings — revenue scale + margin firming against entry trigger; FY27 NDAA markup.

• HEIHEICOdefense

  hold11.88%68
  flow steadyconf

  Funder

  Defense/aero aftermarket compounder — Mendelson family owners since 1957, 20%+ FCF CAGR, disciplined serial acquirer ($1B+/yr). Q2 beat FSG $929M vs $864M, ETG $460M vs $396M; NDAA mandatory funding + commercial aerospace MRO. Textbook 20-30yr franchise. Out of patch scope — held unchanged.

  Unwind

  A value-destroying acquisition, FCF-CAGR break, or sustained commercial-aero MRO downturn. The 54x forward P/E is a valuation watch item near the §6.3 threshold, not a thesis break; lock=1d.

  Catalyst

  HEICO next quarterly print; acquisition-pace cadence.

• PSNdefense

  hold2.01%50
  flow softeningconf

  Funder

  C5ISR/cyber government-services prime, sticky multi-year IDIQ backlog. MDA SHIELD $151B-ceiling IDIQ, DTRA cyber $137M, USAF MATOC $15B, Navy WEXMAC $10B — all prime awards per Parsons IR; $1.1-1.4B Blatnik Bridge win (6/16). Out of patch scope — held unchanged.

  Unwind

  Book-to-bill staying below 1.0 as task-order pace lags IDIQ ceilings, or further confidential-contract softening. Score 50 is below the intact threshold — on a tight leash pending next quarterly book-to-bill; lock=12d.

  Catalyst

  Next quarterly book-to-bill — needs >1.0 recovery to confirm.

• LLYEli Lillybiology

  hold2.50%70
  flow steadyconf

  Funder

  Premier pharma compounder — Mounjaro/Zepbound GLP-1 franchise + retatrutide Ph3 (~70lb loss) extends a multi-decade runway, deep margins, strong capital return. Profitable and cash-generative — insulated from the rate vise. Norway GPFG and large AM holders. Out of patch scope — held unchanged.

  Unwind

  A >10pt US GLP-1 share loss to Novo, a retatrutide Ph3 failure, or a CMS Part D tirzepatide-pricing shock. Lock=27d unlock=none.

  Catalyst

  CMS Part D negotiation updates on tirzepatide; retatrutide Ph3 readouts.

• VRTXVertex Pharmaceuticalsbiology

  hold2.18%76
  flow steadyconf

  Funder

  CF monopoly (Trikafta) cash engine, self-funded, high margins; Casgevy gene therapy + suzetrigine (Journavx) non-opioid pain optionality. Profitable, insulated from the rate vise. GPFG/large AM holders; Casgevy pediatric expansion. Out of patch scope — held unchanged.

  Unwind

  A CF-franchise competitive threat, a suzetrigine commercial flop, or pipeline (CTX611) data failure. Score 76 intact; lock=27d unlock=none.

  Catalyst

  Suzetrigine/Journavx launch trajectory; non-CF pipeline FDA scheduling.

Warnings

• stale_policy guard: MU add → hold (delta zeroed)
• tilt_truncated: biology tilt -5pp truncated at 0% cap
• §2 cap re-applied after pillar tilt: HEI
• position_count: 17 < 18 after mechanical closes (§2 floor) — book under-populated; strategist must open replacements next run
• ledger stale (flagship_performance.json missing) — rebalance diffed against intent

Conviction-lock actions

1. Turn 1ConvictionGuard (§6.5d)

   —
2. Turn 2ConvictionGuard (§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.