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 — 2026-07-23
Headless run (Thu). Research output, not advice — brain never trades. Marks are live 07-23 intraday
twelvedataquotes (markets open); 41/41 symbols priced, 0 429s. Supabase sync + push are performed by the wrapper, not this run.
Top of mind
A thin-vertical chain graduated to a mechanism on primary evidence, and a headline thesis was confirmed by an earnings print the market sold anyway. (1) The AI-gas chain graduated: EQT's Q2 8-K + EIA .gov close the "verify vs a fetched primary" gap held open since 07-22 → new mechanism ai-gas-demand-to-appalachian-producer-price-floor (long EQT/EXE, conviction medium). EQT printed realized $2.65/Mcfe (basis −$0.67 favorable), net debt cut $7.7B→$5.5B, FY volume guide raised, 2027 $4.51 call strikes retaining upside, and signed a 10-yr CPV deal priced to PJM power (an uplift above in-basin) — the price floor being contractually written. Stock +8.5% ($54.01). EIA's own balance (demand>supply near-term, Henry Hub $3.10→$4.00) supports a modest, durable floor, not a squeeze — so the torque is balance-sheet, not gas-price. (2) NOW Q2 beat-and-raise confirmed agentic-ai-seat-erosion-to-saas-rerate first-party — McDermott: "Are we worried about seat compression? Not at all. 50% of our net new business is already non-seat-based"; sub-rev +23% cc, AI ACV crossed $1B→$1.5B, FY guide raised — the Step-4 falsifier did not fire. Yet NOW fell −6.5% ($95.46, −55% off high). The mechanism strengthened while the tape repriced it: the divergence is the story.
Strongest-conviction buys
Ranked shortlist (4c). Prices 07-23 intraday, twelvedata.
| # | Ticker | Causal chain (mechanism) | Conviction | Valuation vs base | Fundamentals (one line) | Next catalyst |
|---|---|---|---|---|---|---|
| 1 | MU | hbm-cowos-as-binding-bottleneck | High (0.80) | Yes — $959.48 (−1.2%, −23% off hi) | ~6× fwd, 84.9% GM, HBM booked thru CY2027 (SemiAnalysis Rubin corroborates HBM4 demand today) | HBM4 ramp / SK Hynix reads |
| 2 | CCJ | ai-capex-to-power-and-materials-cascade + kazatomprom-supply-cut-to-western-uranium-premium | Med-high (0.73) | Yes — $90.37 (+1.9%, −33% off hi) | Live −20.3% abs / −1.1% sector (drawdown is the name) | Utility contracting |
| 3 | TSM | hbm-cowos-as-binding-bottleneck | Med-high (0.70) | At base — $421.21 (−0.8%, −12% off hi) | 18.6× fwd; CoWoS booked thru 2027; Rubin's "simpler tray" eases packaging | N2 ramp |
| 4 | EQT | ai-gas-demand-to-appalachian-producer-price-floor | Medium (0.55, NEW) | Marginal — $54.01 (+8.5%, −21% off hi) | Net debt $5.5B (↓$7.7B), FCF+, FY vol guide raised; retains upside >$4.51 | 2028–2030 demand step; EIA STEO |
| 5 | NOW | agentic-ai-seat-erosion-to-saas-rerate | Med-high (0.62) | Yes — $95.46 (−6.5%, −55% off hi) | Q2 beat+raise, falsifier didn't fire; base $140–150 (~+50%) | Q3 print; ACV conversion |
⚠ LIN (high, 0.78) stays off the buy list — $508.67, −7% off high, "through target." Watchlist. LLY (med-high 0.60) marginal on valuation → watchlist.
Watchlist
| Ticker | Thesis | Conviction | Price | Undervalued vs base? | Next catalyst |
|---|---|---|---|---|---|
| MU | hbm-cowos-as-binding-bottleneck | High | $959.48 | Yes | HBM4 ramp |
| EQT | ai-gas-demand-to-appalachian-producer-price-floor | Medium | $54.01 (+8.5%) | Marginal (ran on print) | Demand step 2028–30 |
| CCJ | kazatomprom-supply-cut-to-western-uranium-premium | Med-high | $90.37 | Yes | Contracting |
| TSM | hbm-cowos-as-binding-bottleneck | Med-high | $421.21 | At base | N2 ramp |
| NOW | agentic-ai-seat-erosion-to-saas-rerate | Med-high | $95.46 (−6.5%) | Yes | Q3 print |
| LLY | glp1-substitution-to-bariatric-surgery-decline | Med-high | $1,163.01 | Marginal | Q2 early Aug |
| INTC | taiwan-chokepoint-to-allied-reshoring | Medium | $102.62 (−2.7%) | ⚠ exit-flag | Q2 TONIGHT 07-23 AMC |
| CME | client-float-interest-to-custodian-exchange-nii | Medium | $249.24 (+5.0%) | No | Q2 (07-23) / rate path |
| MP | china-ree-controls-to-us-producer-stack | Medium | $45.68 (near 52w low) | Yes | Nov 10 deadline |
Undervalued candidates (today)
- agentic-ai-seat-erosion-to-saas-rerate — NOW −6.5% on a beat-and-raise to −55% off high; the Q2 print de-risked the mechanism (falsifier didn't fire) while the tape sold it — the sharpest undervalued-vs-base divergence today.
- china-ree-controls-to-us-producer-stack — MP $45.68, near its 52w low ($43.82), into the Nov 10 China REE deadline.
New chains to investigate (hypothesis-stage)
- ai-gas-egress-buildout-to-appalachian-midstream-tariff-capture — WMB/KMI/ET (prospect-chains, thin energy-ex-AI vertical). The same gas-demand forcing function requires an Appalachian egress build-out whose pipeline operators earn a FERC-regulated, commodity-price-insulated tariff — a more durable beneficiary than the producers. To graduate: cite a midstream operator's Appalachian project ROIC / contracted-capacity book (steps 3-4 ⚠ unverified). → /explore-chain.
New theses (now active)
- ai-gas-demand-to-appalachian-producer-price-floor — EQT/EXE, conviction medium. Graduated from hypothesis on EQT Q2 8-K + EIA primaries. The upstream-molecule leg of the AI-power cascade the book hadn't sized.
Updated theses (existing active)
- agentic-ai-seat-erosion-to-saas-rerate — Q2 confirmed the winner-long fork first-party (non-seat 50% of net-new; AI ACV $1B→$1.5B; beat+raise). Falsifier didn't fire. NOW −6.5% on the print — mechanism strengthened, tape disagrees.
as_of→07-23. - hbm-cowos-as-binding-bottleneck — SemiAnalysis (Vera Rubin) independently corroborates HBM4 (288GB, 2.8× BW) as the binding inference constraint + simpler CoWoS tray for TSM. Evidence-type diversified.
as_of→07-23. - up-nsc-transcontinental-merger-to-pricing-power — timeline reaffirmed (STB supplemental 07-27, mid-2027 close); named coordinated opposition (BNSF/CN motions, CPKC Creel, CSX Angel). UNP Q2 printed 07-23, actuals pending.
Contradicted / weakened
- No active thesis falsified today; 0 calibrations filed. (NOW's beat confirmed its mechanism.)
Open questions worth a human's eye
- ⚠ The 4e calibration gate never fires (6th consecutive mark). Both live flags < −10% sector — PWR −12.3%, TJX −11.0% — are
medium(0.50 / 0.40); 4e only reads high/med-high. Gate-spec calibration filed 07-17; not re-filed. - NOW beat-and-raise, sold −6.5%. The mechanism confirmed but the equity fell — is the market pricing a margin-mix worry (AI GM ~50–60% < seat 80–90%) the thesis flags, or just multiple-normalization? The Q3 ACV-conversion print is the tiebreaker.
- INTC Q2 tonight (07-23 AMC) — two INTC mechanisms sit at −2.5%/−3.1% sector; the print re-tests both. EQT's supply-side falsifier (Permian associated gas 2029–30) is the standing risk to the new gas chain.
What I looked at
- Run status: headless; steps 0.5–7 in-turn; price fetch paced 8/min — 41/41 priced (SSNLF OTC stale $65.21), 0 429s (api.twelvedata.com reachable). PAPER-LEDGER not due (weekly; last 07-17, next ~07-24).
- Breadth (2a): ai-infrastructure 50% ⚠ over → net-new steered to thin verticals: energy ex-AI (gas mechanism + midstream hypothesis). No new AI-infra beneficiary drafted (SemiAnalysis Rubin only corroborated existing HBM/CoWoS).
- Ingestion veins (0.5–0.7): podcast 0 (6 new all filtered; note: an Odd Lots "NY Gov on Data Center Moratorium" ep filtered out — "data center" isn't a keyword); earnings 1 (NOW Q2 — beat+raise; UNP/INTC report today AMC/same-day, SLB 07-24 — not yet available); feeds 1 (SemiAnalysis Vera Rubin, partial/paywalled; FabKnowledge/Apricitas nothing new, Construction Physics roundups only).
- Steps 1–2b: 2 autoresearch (gas-chain graduation-verification → primaries; UNP-NSC transport refresh). prospect-chains: 1 auto-drafted (ai-gas-egress-buildout-to-appalachian-midstream-tariff-capture, WMB/KMI/ET); Meta merchant-vs-custom still held (single-source).
- Sources: 4 promoted + 4 ingested → 2 created (ai-gas-demand-to-appalachian-producer-price-floor mechanism, gina-mastantuono person), 6 updated (agentic-seat-erosion / hbm-cowos / up-nsc mechanisms; eqt / william-mcdermott; gas hypothesis marked graduated), 2 zero-novelty (SemiAnalysis, UNP), 0 contradictions. 4 stale off-topic clippings left in inbox.
- Live feedback (4e): 23 rows, 2 flags < −10% sector (PWR −12.3%, TJX −11.0%), both
medium→ 0 auto-calibrations. Best: CF/CBAM +8.6%, WST +8.9% sector, DNN +11.6%. - Signal feed (7): full active/armed set — see below. +1 net-new row (EQT). Prices re-marked 07-23;
as_ofre-dated only for mechanisms that moved a step today (seat-erosion, hbm-cowos, up-nsc, the new gas row); others carried forward. Supabase sync performed by the wrapper, not this run.