Stock Demon
Overview
Strategy
Portfolio
Earnings
Digest
Health
Architecture
Private cockpit · Not investment advice
Stock Demon
Overview
Strategy
Portfolio
Earnings
Digest
Health
Architecture
Private cockpit · Not investment advice
AI-infra market synthesis. Week ending 2026-07-12. Published 2026-07-14.
Regime. A pullback within an uptrend. About 81% of the universe still sits above its 200-day average — the long AI-infra bull is intact — but only roughly a third remain above their 50-day line. Short-term breadth is thinning, and Monday's tape confirmed it.
Movers (5-day). Up: ANET +8.8% (its 7060XE7 1.6Tbps AI-switch launch plus a round of analyst target hikes), HPE +8.7%, and NBIS +7.9% on continued AI-cloud demand. Down: ORCL −6.9% (bearish-vs-expectations news, and it broke to a fresh 52-week low), SMR −6.7% (also near its 52-week low), and INTC −6.6%.
Risk. Monday 2026-07-13 was a broad risk-off day — the deterioration scan flagged 21 of 42 names across all six layers, four at high severity (ALAB, CRDO, SMR, ORCL). That pattern reads as market-wide de-risking rather than anything name-specific. The model portfolio's drawdown is −12.9% from its peak, still clear of the −20% kill switch.
Portfolio. Monday's rebalance rotated the book: sold AMD (news cooled into event risk), ARM (−32%, through its stop), and ENTG (−24%, through its stop); bought CRDO, MU, and ANET; held ALAB, DELL, EQIX, MRVL, NVDA. One flag worth watching — three of the eight names are now networking (ALAB, CRDO, ANET), a correlated sleeve, and networking was among the layers rolling over Monday. Scorecard since the 2026-06-01 go-live: the model book is at 101.85 vs the equal-weight basket at 91.28, ahead by ~10.6 points — though it's early and this is one short window.
Earnings just landed. Quiet week — no universe names reported.
Next week. A busy stretch opens: ASML (7/15), TSM (7/16), GEV (7/22), then GOOGL, INTC, and DLR all on 7/23. None are currently held. Most sit high in their range going in (more downside asymmetry on a miss); DLR is the least stretched of the group.
Caveats: the mechanical momentum edge is validated but rests on one regime path and is momentum-driven — don't extrapolate. The surprise tilt is a regime-stable commodity signal; news-vs-expectations is unproven-live-only. Not investment advice.