TSMC may join Musk's mega chip factory while Intel bleeds — short Intel or pair it long TSMC

Verdict: Intel-versus-TSMC thesis is credible but untradeable until momentum cools

The thesis has a real asymmetry behind it: Intel's Q2 2026 (ended June 27, 2026) still showed an $11.0B net loss and a net margin of -68.4% even as revenue rose 18.8% quarter over quarter to $16.1B, while TSMC's fiscal 2024 delivered revenue up 33.9%, a 45.7% operating margin, and free cash flow up 197.9% to TWD 870B. Provisional insider filings for the period ending September 30, 2026 lean the same way — about $6.5M of net open-market selling at Intel versus roughly $370K of net buying at TSMC — and Intel has paid no dividend over the trailing twelve months while TSMC's trailing payout sits at $3.85 per share. The strongest argument against the trade is that Intel's turnaround is visibly in motion (free cash flow swung from -$2.5B to +$4.5B and operating margin back to +11.1%), so the short leg's narrative can decay before the thesis pays. On the strategy side, the completed 60-month backtest returned 86.2% over 39 trades with a 43.6% win rate and a 14.0% maximum drawdown, but exits were approximated on daily trigger bars and the 12- and 24-month windows did not evaluate, and no robust parameter setup was established because the sensitivity evaluation exceeded its time budget. Nothing is actionable today: INTC's RSI sits at 59.5 against a 45 trigger and TSM's at 81.9, with INTC's 20-day EMA still above its 50-day EMA and TSM's ADX below 20, so no entry conditions have armed. Verdict: wait for RSI (14) at or below 45 on at least one leg, or Intel insiders flipping to net buying, whichever comes first.

Trade now: no leg is armed — both entries are waiting on the RSI filter

Nothing is actionable today. The strategy enters a long in INTC or TSM only when four conditions line up: a MACD cross below its signal line, the 20-day EMA below the 50-day EMA, RSI (14) at or below 45, and ADX (14) above 20. Right now INTC trades at $119.33 with RSI at 59.5 — about 14.5 points above the 45 trigger — and its 20-day EMA ($112.69) is still above its 50-day EMA ($106.4), so two of the four conditions are unmet even though ADX at 25.4 clears its 20 bar. TSM is further away: at $472.78 with RSI at 81.9, roughly 37 points above the trigger, EMA (20) $444.27 above EMA (50) $431.93, and ADX at 16.6 below its 20 bar. Both momentum legs are near a MACD signal cross, but that alone does not arm an entry. When a leg does trigger, the risk is mechanical. The fixed-risk sizing uses a stop at a 2.4% loss and a take-profit at a 4.8% gain — a 2-to-1 reward-to-risk per position — with positions capped at 25% of the book. Secondary exits also apply: RSI above 52 or 45 days in the trade closes a signal position, and a close below the second support level stops the leg out. For INTC that second support sits at $110; for TSM it is $460. 'Wait' here means concrete things: do not pre-position for the bearish Intel-versus-TSMC thesis. The idea's own argument — that a Terafab decision for TSMC would strip Intel of its most visible foundry prospect — is a narrative, not an entry signal, and the rules only act once price cools enough to put RSI at or below 45 on at least one leg. The completed 60-month backtest for this rule set returned 86.2% over 39 trades with a 43.6% win rate and a 14.0% max drawdown, which is why patience for the exact conditions matters more than urgency. Parameter-sensitivity evaluation exceeded its time budget, so no robust nearby-parameter setup was established — trade the published rules as written.

The bear case has the numbers, the news, and the insiders on its side

The thesis rests on an asymmetry, and the latest filings back it. Intel's most recent quarter (ended June 27, 2026) still shows an $11.0B net loss and a net margin of -68.4%, worse than the prior quarter's -27.5%, even as revenue rebounded 18.8% quarter over quarter to $16.1B. Meanwhile TSMC's fiscal 2024 delivered revenue up 33.9% to TWD 2.89 trillion, a 56.1% gross margin, a 45.7% operating margin, and free cash flow up 197.9% to TWD 870B. The idea argues Intel's whole story depends on winning marquee foundry customers; the fundamentals say the gap is still wide. The catalyst in play is live: per the MarketWatch piece published October 5, 2026, Musk hinted TSMC may join his mega chip venture, and Intel's stock took a hit on the news. That is precisely the scenario the idea is built around — Intel losing its most visible potential foundry customer while TSMC adds another anchor to an AI-driven backlog. Ownership posture leans the same way. Provisional filings for the period ending September 30, 2026 show net open-market insider selling at Intel of roughly $6.5M across 6 reporters, versus modest net insider buying at TSMC (~$370K). Insiders closest to Intel's story appear to be reducing exposure, not adding ahead of a foundry win. Capital-return behavior reinforces who has pricing power. Intel has paid no dividend over the trailing twelve months (its last payment was $0.125 per share in August 2024), while TSMC's trailing dividend is $3.85 per share and 2026 annual payouts are running about 32.8% above 2025. A company that suspended its payout to fund a foundry bet it may lose is a structurally weaker holder of capital than one compounding distributions. On the quant side, the strategy's completed 60-month backtest on daily bars traded 39 times and produced an 86.2% return with a 43.6% win rate and a 14.0% maximum drawdown…

Scores

• Conviction score breakdown: 54
• Thesis support: 75
• Trade readiness: 20
• Risk quality: 50
• Backtest evidence: 55
• Fundamentals trend: 70

Watch items

• INTC — RSI (14)
• INTC — EMA (20) vs EMA (50)
• INTC — Close vs second support
• INTC — Insider open-market net flow
• TSM — RSI (14)
• TSM — ADX (14)
• TSM — Close vs second support
• TSM — Insider open-market net flow