Lam Research just reported earnings confirming equipment makers are seeing massive demand for AI chip manufacturing tools. SanDisk separately beat revenue expectations because data centers are buying memory storage hand over fist. Meanwhile, Bank of Ameri

Verdict: A compelling AI thesis trapped behind a broken entry engine

The fundamental case for AI-driven semiconductor demand is real: Lam Research grew fiscal 2025 revenue 23.7% to $18.4B with a 32% operating margin (97th percentile), while Micron surged 115% to $37.4B at a 26.1% operating margin (94th percentile). But Micron's diluted EPS fell 76.9% year over year despite that revenue explosion, and its history of swinging to negative gross margins as recently as fiscal 2023 raises the question of whether this cycle is closer to its peak than its beginning. The strategy's entry conditions have not triggered once across 1,199 daily bars over 60 months, and the compiled rules contain a logical contradiction that structurally prevents entries; no robust parameter setup was established to resolve this. With LRCX 2.1% above nearest support at $299.32 and MU just 0.8% above $880.23, the setup is waiting for a pullback-to-support reversal that has not arrived. **Conviction breakdown** - Thesis support: 72 — AI demand is confirmed by earnings and analyst calls, though EPS deterioration at MU tempers conviction. - Trade readiness: 15 — Entry rules never triggered across 1,199 bars; no robust setup was established. - Risk quality: 35 — The stop-loss framework (2.4% stop, 4.9% target) is reasonable, but the portfolio's 75% expected max drawdown is extreme. - Trigger proximity: 38 — MU is closest at 0.8% above support; LRCX needs a 2.1% pullback; both need a reversal candle. - Fundamentals trend: 68 — LRCX shows accelerating margins and free cash flow; MU's revenue growth is strong but EPS and margin history are cyclical.

Trade now

**Do nothing today.** Every primary entry condition across this seven-stock semiconductor basket is currently unmet, and several are structurally blocked by contradictory rule logic. The idea argues that Lam Research's earnings, SanDisk's revenue beat, and Bank of America's defense of Micron all confirm AI-infrastructure order books are still filling. That fundamental thesis may be sound, but the compiled entry rules have not produced a single trigger across 1,199 daily bars evaluated over a 60-month window — this is a watch-list setup, not an actionable signal. The most functional entry path is the support-touch rule, which requires a stock's daily low to touch or penetrate the nearest support level while the close finishes back above it. For LRCX at $305.77, nearest support sits at $299.32 — roughly $6.45 below. For MU at $887, nearest support is $880.23, about $6.77 away. Both need a modest pullback to come into range. The Bollinger Band condition is closer: LRCX needs a $11.90 drop to reach the upper band at $293.87, and MU needs a $36.37 decline to its band at $850.63. The primary Bollinger and OBV entries are currently far from triggering on the price conditions and structurally impossible on the self-referential logic (close must be both below negative three and above itself). The OBV condition reads as unknown across all tickers because that indicator is returning null values. The research author has authorized a bounded expanded search to find thesis-consistent thresholds that can produce evaluable history while preserving the intended pullback-to-support setup. Until that revised rule set is available, waiting means monitoring the support levels named above and watching for a reversal candle if price reaches them. If a support-touch entry were to trigger, the stop loss is set at 2.4% below entry and the take-profit target at 4.9% above, yielding an effective reward-to-risk ratio of approximately 2 to 1. No robust parameter setup was established through sensitivity testing; the bounded walk-forward search produced zero supported variants near the current indicator configuration.

The physical AI buildout still shows up in the numbers

The idea's core argument — that AI data center demand is still filling order books for semiconductor equipment and memory…

Scores

• Conviction score breakdown: 46
• Thesis support: 72
• Trade readiness: 15
• Risk quality: 35
• Trigger proximity: 38
• Fundamentals trend: 68

Watch items

• LRCX — Price vs nearest support
• LRCX — Price vs Bollinger upper band
• MU — Price vs nearest support
• MU — Price vs Bollinger upper band
• LRCX — MACD (12,26,9)
• MU — MACD (12,26,9)
• SAND — Price vs Bollinger upper band
• LRCX — RSI (14)
• LRCX — Price below -3
• LRCX — Price below Bollinger (20)
• LRCX — Price above Price