The market is hyper-focused on whether AI chipmakers can justify their sky-high valuations, punishing even good numbers. But the real, guaranteed growth is flowing to the picks-and-shovels companies. Lam Research makes the equipment that physically etches

Verdict: Wait for the washout — strong thesis, dormant setup

The picks-and-shovels thesis is fundamentally sound: EMCOR and Quanta collectively hold $70B in backed-up work orders, and Lam Research grew revenue 23.7% year-over-year to $18.4B with a 97th-percentile operating margin of 32.0%. But the mean-reversion strategy designed to exploit this idea is nowhere near firing. All three tickers carry RSI readings in the low-to-mid 60s against an entry requirement below 40, with EME at $827.57 needing a drop to $766.39, PWR at $691.51 needing $644.86, and LRCX at $312.08 needing roughly $294.78. The single completed backtest trade on LRCX returned 5.1% over 24 months, yet the 60-month evaluation produced zero triggers at all — meaning this setup demands a rare coordinated oversold washout that may not recur for years. No robust parameter setup was established; the sensitivity evaluation exceeded its time budget without producing a recommendation. The 0.78 correlation between EME and PWR means a triggering selloff would effectively hit two-thirds of the basket as one concentrated infrastructure bet, while LRCX diversifies with near-zero correlation. ### Conviction Breakdown | Dimension | Score | |---|---| | Thesis support | 82 | | Trade readiness | 15 | | Risk quality | 38 | | Backtest evidence | 20 | | Fundamentals trend | 86 |

Trade now: wait for a pullback, not a chase

**All three tickers are in wait mode.** The entry rules require price at or below the 50-day EMA, price at or below the lower Bollinger band (20-period, 2 std dev), RSI (14) below 40, and a cross back above 30 — a combination that asks for a sharp oversold washout before buying. Right now, none of those conditions are close to firing. EME sits at $827.57 versus an EMA-50 of $793.84 (needs to drop about $34); PWR is at $691.51 versus $664.11 (needs about $27 lower); LRCX is the furthest from its trigger, at $312.08 versus $293.39 EMA-50 and a $294.78 lower Bollinger band (needs roughly $17–19 of downside). RSI readings of 66.4 (EME), 64.9 (PWR), and 61.0 (LRCX) all need to fall below 40, meaning each stock requires a 21–26 point RSI deterioration before the setup is active. **What "wait" means concretely:** do not initiate positions today. Set price alerts at each ticker's lower Bollinger band — $766.39 for EME, $644.86 for PWR, $294.78 for LRCX — and plan to act only if RSI simultaneously pushes below 40 and then crosses back above 30. The stop is hard-coded at a 2.3% loss from entry, with profit targets at 4.7% or a Fibonacci 127.2% extension, producing an effective reward-to-risk ratio of roughly 2:1 on the percentage-based exits. The backtested sample that triggered on LRCX delivered a 5.1% return with a 9.5% max drawdown over 24 months, though the strategy generated zero triggers across the longer 60-month evaluation window for that pair — the rules are highly selective and require patience. **Risk parameters to note:** position sizing uses a fixed-risk method of 2.3% per trade, with a maximum 20% allocation per position. The stop loss at 2.3% is tight; entries are designed to catch a bounce from extreme oversold conditions, and a failure to bounce quickly invalidates the thesis. No parameter optimization is recommended — sensitivity analysis exceeded its time budget without identifying a superior nearby setup, so the published rules stand as-is. **Diversification context:** EME and PWR carry a 0.78 correlation, so they tend to move together. LRCX is essentially uncorrelated with both (correlations of −0.03 and −0.02), giving the basket a diversification ratio of 1.47. This means that if a broad selloff triggers entries, LRCX may fire independently of the electrical-infrastructure pair.

The Picks-and-Shovels Thesis Is Backed by Hard Numbers

The idea's central argument — that AI infrastructure spending flows to picks-and-shovels companies with locked-in revenue — is strongly supported by the fundamentals across all three tickers. Per the Yahoo Finance backlog piece, EMCOR and Quanta collectively report $70 billion in combined backlog, providing tangible evidence of the multi-year revenue visibility the thesis depends on. That backlog is not merely top-line: EMCOR converted it into $1.19B in free cash flow this fiscal year, placing it in the 97th percentile of its Industrials peer group. Quanta generated $1.62B in free cash flow, ranking even higher at the 97.6th percentile. Lam Research anchors the semiconductor-equipment leg with the strongest margins in the basket. Its 48.7% gross margin and 32.0% operating…

Scores

• Conviction score breakdown: 48
• Thesis support: 82
• Trade readiness: 15
• Risk quality: 38
• Backtest evidence: 20
• Fundamentals trend: 86

Watch items

• LRCX — Price vs lower Bollinger band
• LRCX — RSI (14)
• EME — Price vs EMA (50)
• EME — Price vs lower Bollinger band
• EME — RSI (14)
• PWR — Price vs lower Bollinger band
• PWR — RSI (14)
• EME — MACD (12,26,9)
• PWR — MACD (12,26,9)
• LRCX — MACD (12,26,9)
• EME — Price
• EME — Price
• EME — RSI (14) below 40
• EME — RSI (14) crossed above 30