Stock-market dispatch — 2026-07-29
Headless run (Wed). Research output, not advice — brain never trades. Marks are live 07-29 intraday
twelvedataquotes; 44/44 symbols priced, 0 429s (SSNLF OTC excluded). Step 4d (weekly ledger) not due (Fri cadence; last mark 07-24). Supabase sync + push are performed by the wrapper, not this run.
Top of mind
The rotation the ROI-reckoning thesis predicted showed up on the tape — and the day's fresh primary material confirmed the mechanism. AI semis were wrecked (MU −8.9% $820.53, ARM −8.1%, DELL −8.2%, INTC −5.9%, PWR −5.2%) while defensives/value rallied (BLK +3.3%, TMO +3.0%, ROST +3.2%, TJX +2.8%, UNH +2.7%, LLY +1.9%) — a clean instance of ai-roi-reckoning's "market stops rewarding the spenders" tell. Two sources hardened it: The Compound's michael-batnick named the reaction-function flip (Alphabet −6% on a good quarter — "the only thing investors focus on is CapEx and free cash flow, one's going higher, the other's going lower… punish at the open"), and the megacap-capex autoresearch found the actual trigger has moved to credit — record CDS demand on ~$100B of AI-capex bonds and a ~$1.5T debt need — even as capex still accelerates (>$700B 2026 big-four, +77%; >$1T 2027). So the derate is a financing/valuation event, not a demand crack (Google backlog +$50B, MSFT $627B backlog intact). Correctly, no net-new AI chain was drafted (breadth ai-infra 47% ⚠ over); the day's net-new went to a thin materials vertical (line-pipe/OCTG off the gas-egress buildout), and the ingest added a policy accelerant (Sen. Armstrong's permitting-reform bill) to the egress chain.
Strongest-conviction buys
Ranked shortlist (4c). Prices 07-29 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 — $820.53 (−8.9%, −35% off hi) | ~6× fwd, HBM booked thru CY2027; the semis rout = better entry; MSFT's $25B "component-pricing" capex today = fresh memory-inflation corroboration | HBM4 ramp / SK Hynix reads |
| 2 | CCJ | ai-capex-to-power-and-materials-cascade + kazatomprom-supply-cut-to-western-uranium-premium | Med-high (0.73/0.52) | Yes — $86.96 (−2.7%, −36% off hi) | Uranium premium intact — NXE +8.1% / DNN +7.1% legs beat sector; CCJ the liquid anchor | Cameco Q2 (07-31) |
| 3 | NOW | agentic-ai-seat-erosion-to-saas-rerate | Med-high (0.62) | Yes — $110.62 (+4.8%, −45% off hi) | Q2 beat+raise held; base $140–150 (~+35%); 3rd up-leg since beat = the gap-close the thesis flagged | Q3 print; ACV conversion |
| 4 | TSM | hbm-cowos-as-binding-bottleneck | Med-high (0.70) | At base — $392.31 (−1.7%, −18% off hi) | 18–19× fwd; CoWoS booked thru 2027; +custom-ASIC broadening of N4/N5 | N2 ramp |
| 5 | LIN | ras-laffan-halt-to-lin-helium-pricing-power | High (0.78) | Marginal — $511.18 (+0.8%, −7% off hi) | 30%+ margin; caught the defensive bid today; near target | Helium contract reads |
⚠ LLY (med-high 0.60) — $1220.66 (+1.9%), at 52w hi → priced-for-perfection, watchlist. SLB (med 0.52) — $49.98 (−3.0%), gave back more of the print pop → watchlist. BLK (med-high 0.62) — $1097.55 (+3.3%), the DB→DC bid caught today's rotation, near target → watchlist.
Watchlist
| Ticker | Thesis | Conviction | Price | Undervalued vs base? | Next catalyst |
|---|---|---|---|---|---|
| MU | hbm-cowos-as-binding-bottleneck | High | $820.53 (−8.9%) | Yes | HBM4 ramp |
| CCJ | kazatomprom-supply-cut-to-western-uranium-premium | Med-high | $86.96 | Yes | Q2 07-31 |
| NOW | agentic-ai-seat-erosion-to-saas-rerate | Med-high | $110.62 (+4.8%) | Yes | Q3 print |
| TSM | hbm-cowos-as-binding-bottleneck | Med-high | $392.31 | At base | N2 ramp |
| EQT | ai-gas-demand-to-appalachian-producer-price-floor | Medium | $51.66 (−0.7%) | Marginal | 2027 floor; DC-isolation |
| INTC | tsmc-saturation-to-intel-anchor-stack | Medium | $86.30 (−5.9%) | ⚠ external-anchor unproven | 14A firm decisions H2'26 |
| SLB | slb-oilfield-pivot-to-dc-infrastructure | Medium | $49.98 (−3.0%) | No | Q3 DC run-rate |
Undervalued candidates (today)
- hbm-cowos-as-binding-bottleneck — MU $820.53, −8.9% in a broad semis rout (ARM/DELL/INTC all −6 to −8%) on a thesis whose supply-constraint leg keeps getting first-party corroboration (MSFT's $25B component-pricing capex today); ~6× fwd, −35% off high.
- agentic-ai-seat-erosion-to-saas-rerate — NOW $110.62, +4.8% and rising against a red tape, still −45% off high, base $140–150 — the market continuing to close the seat-erosion gap the thesis flagged.
New chains to investigate (hypothesis-stage)
- gas-egress-buildout-to-line-pipe-octg-supplier — TS / NUE / STLD. Prospect-chains (2b): the confirmed ~20 Bcf/d Appalachian/LNG-feedgas egress build-out consumes large-diameter line pipe (+OCTG) — the material input one link upstream of the midstream toll-road, in the absent materials cluster (breadth 2a). To graduate: a pipe-maker's North-American line-pipe order-book tied to this build-out (vs the broad drilling cycle) — steps 2–3 ⚠ unverified. → /explore-chain. priority low.
New theses (now active)
- None. Corroboration/refresh day — no new mechanism graduated; net-new was steered to a thin-vertical hypothesis (per breadth 2a).
Updated theses (existing active)
- ai-roi-reckoning — reaction-function flip now confirmed at the megacap level. michael-batnick/josh-brown (The Compound): Google −6% on a good quarter, capex↑/FCF↓ "punish at the open"; MSFT $627B backlog = the patience. + megacap-capex autoresearch: >$700B 2026 capex but the raise sells off, and the trigger is a credit break (record CDS on ~$100B AI bonds, ~$1.5T debt). No step-status flip.
as_of→07-29. - mega-issuance-peak-to-ai-capex-derate — Step 4 now has a machine-checkable financing trigger (CDS spreads on AI-capex issuers + the $1.5T issuance pace) that fires before any vacancy signal. Held low-med.
as_of→07-29. - ai-gas-demand-to-appalachian-producer-price-floor — EXE Q2: realized $2.90/Mcf (NE Appalachia $2.15 — basis intact), no 2027 guidance / no DC-vs-LNG color; CEO framed on the $1.25B Twin Eagle marketing buy → the DC-floor is not isolable from actuals; near-term upside stays LNG-led. Reinforces the "floor not yet realized" read. Held medium.
as_of→07-29. - ai-gas-egress-buildout-to-appalachian-midstream-tariff-capture — permitting-reform accelerant. alan-armstrong (R-OK, ex-Williams CEO): the D.C. Circuit vacated Williams' finished, operating Regional Energy Access certificate — the litigation risk his American Energy and Mineral Infrastructure Act of 2026 would end. A conditional catalyst (bullish egress timing if it passes; legislative and uncertain). New catalyst added.
Contradicted / weakened
- No active thesis falsified. 0 new calibrations. 4e gate: 3 flags <−10% sector (PWR −21.4%, TJX −11.5%, UUUU −10.5%) but all three are medium tier → no gate fires. UUUU's label already downgraded to medium (0.52); same vehicle-selection miss (NXE +8.1% / DNN +7.1% legs beat sector) calibrated 07-27 — no re-file.
Open questions worth a human's eye
- Does the credit market fire before the tape? With AI-bond CDS demand at records and ~$1.5T of issuance ahead, the datable trigger for mega-issuance-peak-to-ai-capex-derate may be a spread widening, not a vacancy print. Watch AI-issuer CDS + the pace of new AI-capex bonds.
- Is today's semis rout the ROI-reckoning rotation or a pre-earnings de-risk? MSFT/META (07-29 pm) and AMZN/AAPL (07-30 pm) report into it — the postmortem is the test of whether the "punish the raise" reaction generalizes.
- The gas chain's DC-isolation problem persists — EXE's Q2 disclosed nothing that separates a data-center-driven floor from the LNG-led lift; the next primary read is any Appalachian producer that discloses DC-attributable realized price.
What I looked at
- Run status: headless; steps 0.5–7 in-turn; price fetch paced 8/min — 44/44 priced, 0 429s. Step 4d not due (Fri cadence). Per-concept bull/base/bear narrative carried forward (headless budget); load-bearing marks in the feed + tables above.
- Breadth (2a): ai-infrastructure 47% ⚠ over (43 distinct chains) → net-new steered to the absent materials cluster. Macro buckets rotated to #8 healthcare + #9 financials (thin verticals).
- Ingestion veins (0.5–0.7): podcast 2 (Columbia Energy/Armstrong; The Compound megacap-preview — diarized AssemblyAI, ~$0.33); earnings 0 (no tracked name reported un-ingested — PG's call is today 8:30am ET, transcript not yet out; CCJ/LIN 07-31, PWR 07-30, LLY 08-05); feeds 0 (SemiAnalysis nothing new; Fabricated Knowledge "Close but Not Quite" fully paywalled; Construction Physics off-topic/roundups; Apricitas dormant since May → flag for
pausedreview). - Steps 1–2b: 3 autoresearch (EXE Q2 gas floor; megacap Q2 capex→ROI repricing; thin-vertical healthcare+financials scan). prospect-chains: 1 auto-drafted (gas-egress-buildout-to-line-pipe-octg-supplier; GLP-1/WST + midstream/WMB + credit candidates all deduped against existing pages — a healthy well-connected-wiki signal).
- Sources: 5 promoted + 5 ingested → 1 created (alan-armstrong person) + 1 hypothesis (2b); ~6 updated (ai-roi-reckoning, mega-issuance-peak, ai-gas-demand, ai-gas-egress, glp1-injectable; +quotes batnick/brown). 0 new mechanisms (corroboration day). 1 zero-novelty thesis-grade source (The Compound megacap-preview banter).
- Live feedback (4e): 23 scored rows; 3 flags <−10% sector all medium tier → no gate. 0 new auto-calibrations. Best: NXE +8.1%, DNN +7.1%, STVN +4.7%, CF/CBAM +4.5%.
- Signal feed (7): full active/armed set (see below). Prices re-marked 07-29;
as_ofre-dated only for the rows whose research moved today.