Corporate spending is shifting to power, defense and supply chains — buy the industrial capex winners on dips

Verdict: the capex thesis is earning its keep — wait for the dip, don't chase the strength

The strongest case for this trade is the fundamental confirmation: Caterpillar's Q2 2026 revenue rose 18.0% sequentially to $20.5B with free cash flow up 231.4% to $3.8B, Quanta's free cash flow jumped 403.6%, and Lockheed swung from negative $291M to positive $2.9B — a broad, same-quarter cash inflection that matches the thesis. The strongest case against is the ownership signal: provisional filings as of the September 30, 2026 report period show net open-market insider selling of roughly $123.2M at PWR and $113.6M at CAT — the basket's two largest risk allocations — while CAT, ETN, GEV and PWR all trade with RSIs between 56 and 72, meaning four of five entries need a pullback that has not arrived. The 24-month backtest (46 trades, +62.3%, 39.1% win rate, 7.3% max drawdown) is encouraging but was run on the CAT pair only, no robust parameter setup was established, and exits were filled on daily bars, which the fidelity note flags as possibly flattering results. GEV's eye-catching 50.8% Q1 net margin rests on figures the source flagged as currency-unsupported, and Eaton's Q2 margins contracted. The verdict: the thesis is real, but the disciplined move is to wait at the defined pullback levels rather than chase strength. What would flip this to a buy: a genuine dip where RSI reaches 40 or below and price closes back above the lower Bollinger band, ideally alongside the next ownership refresh showing insider selling cooling.

Trade now: LMT is live, the other four are waiting for their levels

Do not chase today. Of the five names, only **LMT** has its entry conditions live: its 14-day RSI is 19.7, well below both the 50 and 40 thresholds, and its close of $505.51 sits 3.5% below the lower Bollinger band at $523.61 — a bounce back above that band is the trigger to watch. One caveat: LMT's ADX is already at 39.5, above the 35 exit threshold, so any position taken here would start with a signal-based exit condition already active. The strategy also caps risk at a 2.0% stop and a 4.0% take-profit. Note that the backtest evidence for this setup comes from the CAT pair — 46 trades, a 39.1% win rate, +62.3% return, and a 7.3% max drawdown over 24 months — so treat the other names as thesis-consistent extensions rather than independently validated legs. For the other four, the entries are a pullback away. CAT's RSI is 60.6 (it needs to be at or below 50, and at or below 40 for the stricter rule), with the close of $836.33 sitting $24.00 above the lower band at $812.33. ETN's RSI is 56.1, with price about $9.99 above its band at $424.42. GEV at $980.09 trades $34.99 above its $945.10 band. PWR is the furthest extended: RSI at 72.0 and $39.25 above its $640.04 band. "Wait" here means a concrete set of levels: RSI dropping to 40 or lower and a close back above the lower Bollinger band, or a low that touches the nearest support with a close above it (CAT's nearest support is $839.17, ETN's is $430, GEV's is $981.71, PWR's is $670). Sizing is fixed-risk: 2.0% of capital at risk per position, stops placed off the rank-2 support level, and a hard 20% maximum position size. Given the strategy's own history of 39% winners, the math depends on the 4% take-profit versus the 2% stop — that 2:1 payoff is what makes a sub-40% hit rate work. The CNBC framing in the idea is right that pullbacks are the entry mechanism here; the levels above are exactly where those pullbacks convert into positions. One structural note for GEV: the 60-month backtest window could not be completed because of a gap in GEV's daily market history, so the 24- and 12-month CAT windows are the validated evidence. If GEV is your preferred leg, accept that its entry rules are the same code but its own pair was never fully backtested.

The capex winners are actually earning their stripes — the backtest agrees

The thesis is that corporate money is rotating into power, supply chains, and defense — and the most recent reported fundamentals back that rotation with real numbers, not just narrative. Caterpillar's Q2 2026 (period ended June 30) shows revenue of $20.5B, up 18.0% quarter over quarter, with net income up 41.0% to $3.6B and free cash flow jumping 231.4% to $3.8B. Quanta Services posted a 21.4% sequential revenue gain to $9.6B and doubled net income to $451.4M. On GE Vernova, the reported Q1 2026 net margin was 50.8%, up 17.4 percentage points from the prior quarter. These are exactly the profit expansions you would expect if the capex boom is real. The breadth claim also holds in the peer rankings: against roughly 729 industrial peers, CAT sits in the 78th percentile for revenue growth and the 78th percentile for operating margin, PWR ranks in the 81st percentile for revenue growth, and CAT and LMT both sit above the 99th percentile for free cash flow. Five different names, all generating top-decile cash flow or top-quartile growth — profit growth genuinely is broadening beyond chipmakers. The cash-flow proof points matter for the position-trade horizon. LMT swung from negative $291M free cash flow in Q1 2026 to positive $2.9B in Q2, and PWR's free cash flow rose 403.6% to $864.5M. These are not one-off beats but sequential inflections across the group within the same quarter. The completed 24-month backtest supports the trading framing. Over that window the rules produced 46 trades with a 62.3% cumulative return, a maximum drawdown of 7.3%, and a 39.1% win rate — a profile where the losers are cut fast (a 2% stop is in the rule set) and the trend exits do the heavy lifting. The 12-month window also traded, with 4 trades, a 13.0% return, a 75% win rate, and a 5.9% drawdown. Note that no robust parameter setup was established: the sensitivity evaluation exceeded its time budget, so the published configuration is used as-is rather than a tuned variant. The demand-side context also checks out per the cited news: the Bloomberg piece on SK Hynix's $26.5B US listing (October 1, 2026) is direct evidence that mega-scale capital is still being raised for AI-adjacent infrastructure, and UBS's Lovell (October 2) explicitly frames power, supply chains, and defense as the funding recipients. The names in this basket are those recipients.

Insiders are selling, margins at two names slipped, and the pullback risk is the plan's whole premise

Start with the ownership posture, because it is unambiguous. In provisional ownership data as of the September 30, 2026 report period, four of the five names show net open-market insider selling: roughly $123.2M at PWR, $113.6M at CAT, and smaller amounts at ETN and LMT. GEV shows only a single reporter, so its ownership read is thin. The executives running the capex boom are monetizing stock at scale — management pay figures (e.g., GEV CEO Strazik at $6.3M, LMT CEO Taiclet at $9.8M) suggest insiders are hardly underpaid, so the selling is a deliberate distribution signal worth weighing. The fundamental picture is not uniformly strong either. Eaton's Q2 2026 (June…

Scores

• Conviction score breakdown: 59
• Thesis support: 70
• Trade readiness: 35
• Risk quality: 60
• Backtest evidence: 55
• Fundamentals trend: 75

Watch items

• LMT — RSI (14)
• LMT — ADX (14)
• CAT — RSI (14)
• CAT — Close vs lower Bollinger band
• ETN — RSI (14)
• GEV — Close vs lower Bollinger band
• PWR — RSI (14)
• PWR — Insider net open-market flow