2026-07-282026-07-272026-07-242026-07-232026-07-222026-07-212026-07-202026-07-172026-07-162026-07-152026-07-142026-07-102026-06-292026-06-192026-06-182026-06-172026-06-162026-06-152026-06-122026-06-112026-06-102026-06-092026-06-082026-06-052026-06-042026-06-032026-06-022026-06-012026-05-302026-05-292026-05-282026-05-272026-05-252026-05-222026-05-212026-05-202026-05-192026-05-182026-05-152026-05-11
Stock-market dispatch — 2026-06-29
Headless Pi catch-up run (06-20→06-28 gap). Summary of what changed in the wiki; the wiki is the truth. Research output, not advice — brain never trades.
Top of mind
Micron's June-24 print didn't just beat — it converted the #1 chain from "sold-out narrative" into a first-party-confirmed, contractually-floored structural-undersupply thesis, and the day's other big move is that the power-bottleneck cascade went confirmed-grade across three independent source types. MU posted $41.5B rev (+346% YoY), 84.9% GM (beat the ~81% guide), $25.11 EPS, and — the load-bearing part — 16 take-or-pay SCAs (~$100B RPO, $22B deposits, 5-yr floor/ceiling bands), with HBM "fully booked through CY2027 into 2028" and "no line of sight on when supply catches demand" (2026-06-24-earnings-mu-q3-fy2026). hbm-cowos-as-binding-bottleneck's falsifier is re-dated to a machine-checkable kill-switch (CXMT HBM wafer share >12% by 2028 + SCA-discipline-break + AI-demand air-pocket → 2028-29 oversupply). Separately, ai-capex-to-power-and-materials-cascade is now confirmed (SemiAnalysis grid post + Cembalest + Alden all independently say power, not chips, is the binding constraint; nat-gas is the bridge fuel), opening a behind-the-meter chain to Bloom Energy (BE +10.6%) / GE Vernova (GEV +5.6%). Tape: TSM +4.8%, INTC +2%, and space peers ripped (RKLB +16%, ASTS +20%) — contradicting the spacex-peer-fade thesis.
Strongest-conviction buys
Ranked shortlist (4c). Prices 2026-06-29 twelvedata (56/60; ASE/BESI 403, GNKG/HXSCL carry).
| # | Ticker | Causal chain (mechanism) | Conviction | Valuation vs base | Fundamentals (one line) | Next catalyst |
|---|---|---|---|---|---|---|
| 1 | MU | hbm-cowos-as-binding-bottleneck | High (0.80) | At base — $1119.87 −1.1% (digesting +14.6% print pop) | $41.5B rev +346%, 84.9% GM, $18.3B FCF, ~10× fwd P/E (ASP-driven, not a bubble) | SCAs ramp; supply "tight beyond CY2027" |
| 2 | CCJ | kazatomprom-supply-cut-to-western-uranium-premium + ai-capex-to-power-and-materials-cascade | Med-high (0.73) | Yes — $103.59 −0.9%, −23% from hi | Term-U $90 deficit; power-cascade now confirmed-grade; live +5.7% sector | Utility contracting; Vistra/AWS co-location corroborates |
| 3 | INTC | taiwan-chokepoint-to-allied-reshoring + us-industrial-policy-tariff-shield | Med-high (0.68) | Momentum — $130.84 +2.0% | Tariff-shielded US foundry; live ledger-best +15.7% sector | 18A yield / Apple-deal |
| 4 | LIN | ras-laffan-halt-to-lin-helium-pricing-power | High (0.76) | At base — $510.10 −1.8% (near 52w hi) | Net-cash industrial gas, 85-90% contracted; MU print did not touch helium leg | APD Q3 late-July; Ras Laffan restart |
| 5 | WST | glp1-injectable-supply-chain-bottleneck | Med-high (0.66) | Marginal — $352.29 +0.4% (AT 52w hi) | GLP-1 ~18% sales, per-unit consumable, price-agnostic | 503B comment-close ~Jun 29; Medicare $50 demo Jul 1 |
Watchlist
| Ticker | Thesis | Conviction | Price | Undervalued vs base? | Next catalyst |
|---|---|---|---|---|---|
| MU | hbm-cowos-as-binding-bottleneck | High | $1119.87 ↓1.1% | At base (confirmed) | SCAs ramp |
| CCJ | kazatomprom-supply-cut-to-western-uranium-premium | Med-high | $103.59 | Yes (−23% from hi) | Contracting |
| INTC | taiwan-chokepoint-to-allied-reshoring | Med-high | $130.84 ↑2.0% | Momentum | 18A / Apple |
| LIN | ras-laffan-halt-to-lin-helium-pricing-power | High | $510.10 | At base | APD Q3 |
| WST | glp1-injectable-supply-chain-bottleneck | Med-high | $352.29 | Marginal | 503B Jun 29; Medicare Jul 1 |
| TSM | hbm-cowos-as-binding-bottleneck | Med-high | $453.02 ↑4.8% | Near hi (MU read-through) | CoWoS/HBM4 |
| CEG | pjm-capacity-prices-to-nuclear-premium | Medium | $259.79 ↓1.6% | Dislocated; live −13% sector | FERC; co-location |
| CF | hormuz-nitrogen-supply-shock-to-cf-risk-premium | Medium (de-esc) | $104.63 | Lagged food-cascade | H2 nitrogen |
| MP | drone-warfare-demand-to-mp-hree-rerate | Medium | $54.75 ↑1.6% | Yes (−45% from hi) | HREE circuit H2'26 |
| NOW | agentic-ai-seat-erosion-to-saas-rerate | Medium (winner-long) | $100.20 ↑1.9% | Yes | Q2 earnings |
| BE/GEV | btm-onsite-generation-to-bloom-fuelcell-gev-turbine | Hypothesis | $278.75 ↑10.6% / $1103 ↑5.6% | New chain ripping | BTM order backlog |
Undervalued candidates (today)
- CCJ $103.59 — −23% from hi; power-cascade confirmed-grade + term-U $90 deficit; best live cluster (critical-minerals +1.4%).
- MP $54.75 — −45% from hi; REE breadth-thin leg.
- CEG $259.79 — dislocated (live −13% sector) but Cembalest flags Western-nuclear execution-cost gap as the falsifier.
New chains to investigate (hypothesis-stage)
- btm-onsite-generation-to-bloom-fuelcell-gev-turbine — BE/GEV. Grid headroom turns negative → behind-the-meter self-generation; dual-sourced (SemiAnalysis + Cembalest), BE +10.6% today. Distinct from genset CAT/CMI. Graduate via /explore-chain on BE/GEV order-backlog.
- ai-terminal-value-uncertainty-to-credit-curve-steepening — short 30yr CRM/ORCL credit (Barbalat, first-party CIO). Thin financials vertical. Graduate on actual long-dated tech-IG steepening.
- frontier-lab-distrust-to-hyperscaler-gatekeeper-toll — AMZN/MSFT/GOOGL long / neocloud short (Sacks WH readout, bias-discounted; strong Friedberg disaggregation falsifier).
New theses (now active)
- None promoted to active concept — the day's movement was a confirmation (MU) + thin-vertical hypotheses awaiting validation.
Updated theses (existing active)
- hbm-cowos-as-binding-bottleneck — CONFIRMED first-party by MU; ASP-not-volume mechanism + AAPL inflation pass-through (Step 6) added; falsifier re-dated to CXMT 2028 kill-switch.
- ai-capex-to-power-and-materials-cascade — → confirmed (3 source types); nat-gas-bridge + GEV/BE/VST beneficiaries; overbuild + Western-nuclear falsifiers.
- p-and-c-insurer-float-income-rate-regime — first-party CIO confirmation (Barbalat) → priority medium-high.
- ipo-comeback-public-market-value-capture — net equity supply positive first time since 2021 + SK Hynix ~$30B US listing.
Contradicted / weakened
- spacex-ipo-comp-anchor-to-space-peer-fade — fade leg CONTRADICTED by the tape (RKLB +16%, ASTS +20% — halo lifting peers); conviction low-med→low; calibration logged (
thesis-direction-assumed-from-one-print). - iran-fuel-shock-consumer-bifurcation — Hormuz reopen (Sacks MOU + Alden 60-day fragility) weakens fuel leg further; armed-not-closed. Live TJX −10.9% sector (medium conviction → no auto-calibrate).
- Calibrations (2 + 1 follow-up): SpaceX peer-fade contradicted; MU thesis confirmed (
conviction-warranted-by-chain-depth, positive); Warsh higher-for-longer refined to no-cut/no-hike (oil 120→70 removes hike trigger).
Open questions worth a human's eye
- Does the take-or-pay SCA floor genuinely cycle-proof MU, or just relocate risk to the ~80% of DRAM bits not under SCA if hyperscaler capex pauses?
- BTM-power (BE/GEV): is the +10.6%/+5.6% move the start of an order-backlog re-rate, or overbuild front-running (Cembalest "takes the under")?
- Space peers: is the comp-anchor a durable peer support, or does the lockup-unlock still bring the fade?
What I looked at
- Run status: headless Pi catch-up (06-20→06-28 gap); steps 0.5–7 completed synchronously in-turn; price fetch paced 56/60 + 8 ETFs, 0 rate errors.
- Macro buckets: thin-vertical material came via ingest (Cembalest = energy ex-AI; Alden = financials/macro; GameStop = consumer) rather than net-new bucket autoresearch — ai-infrastructure 50% ⚠ over per 2a.
- Ingestion veins (0.5–0.7): podcast 10 transcripts (Bg2 failed — AssemblyAI-400, 8th consecutive; off-scope Dwarkesh/Machiavelli + 4.5hr Acquired/Disney + Capital-Allocators replay skipped); earnings 1 (MU Q3 — the headline); feed 1 (SemiAnalysis grid).
- Step-1: 1 gap-fill autoresearch (MU SCA cycle-break vs cycle-top). prospect-chains: 0 auto-drafts (the 1 novel candidate, Bloom fuel-cells, is over-cluster ai-infra → reported speculative; landed via 4a as hypothesis).
- Sources promoted 13; ingested 13 → 11 new entities (5 company: BE/VST/GEV/CXMT/EBAY · 6 person: Mehrotra/Sadana/Murphy/Cembalest/Alden/Barbalat); 3 new hypothesis chains; 2 core mechanisms upgraded to confirmed; 4 zero-novelty (Dwarkesh/GameStop/Micron-preview/Edelman/Williams).
- Valuation: 56/60 twelvedata (prices into the conviction/watchlist tables + signal feed; ASE/BESI 403, GNKG/HXSCL/SSNLF/CBRS carry).
- PAPER-LEDGER (4d): weekly mark DONE 06-29 (11-day catch-up) — per-position +5.5% sector-excess (ex-KLAC), conviction ordering HOLDS (med-high +14.8% > high +9.7% > medium −7.4%); MU the standout (+33.6% sector, the June-24 gate fired); KLAC still split-anomaly-excluded.
- Live feedback (4e):
performance-2026-06-29(22 pos) — 1 flag < −10% sector (TJX −10.9%) but medium-conviction → 0 auto-calibrations; energy-oil worst cluster (−5.4%), INTC live-best (+15.7%). Calibrations: 2 + 1 follow-up. Conviction buys ranked: 5. New hypothesis chains: 3. - Signal feed: full active/armed set, gated ≥0.4 — see step 7 (Supabase sync performed by the wrapper, not this run).