IBM crash triggers tech panic but broader market is green — buy the semiconductor dip

Verdict: compelling thesis, but wait for the trigger

This semiconductor mean-reversion setup rests on a compelling dislocation: NVDA grew revenue 65.5% year-over-year with a 71.1% gross margin and TSM expanded 31.6% with operating margins in the 99th percentile of tech peers, yet per the Bloomberg piece these names are being dumped in a guilt-by-association rout tied to IBM's crash rather than their own fundamentals. The backtest is genuinely impressive at face value — 176.5% cumulative over 60 months with a 56.9% win rate — but it conceals a punishing reality: the first walk-forward fold lost 13.0% with a 32.6% drawdown and only a 37.5% win rate, meaning the strategy has a demonstrated history of systematically failing in hostile regimes before recovering. The recommended RSI period 17 variant delivered 43.3% on the final 12-month holdout with a 60.7% win rate and a 10.4% max drawdown, which is the cleanest read on how this setup performs in conditions resembling today's. Critically, no entry conditions are live across NVDA, SMH, or TSM right now — NVDA's RSI sits at 55.4 against a 35 trigger — so this is a watchlist setup awaiting its moment, not a trade to force. The conviction is moderate-to-high: the quality of the underlying assets is exceptional, the logic is sound, but the 2.5% stop is razor-thin for names that can gap on earnings, and the basket is one thematic bet wearing three tickers. **Conviction breakdown** - **Thesis support (72/100):** The guilt-by-association narrative is well-grounded in the fundamentals gap, but Bloomberg's framing of the rout as putting earnings in the spotlight cuts both ways. - **Trade readiness (38/100):** All three tickers are well outside entry conditions. NVDA needs a 20-point RSI drop; TSM is closest but ADX at 12.4 is far from triggering. - **Risk quality (45/100):** The 2.5% stop against earnings-gap exposure is tight, and walk-forward fold 1 showed a 37.5% win rate — below breakeven for a 2:1 reward-to-risk structure. - **Backtest evidence (68/100):** A 60-month return of 176.5% with a 56.9% win rate is strong, but the 32.7% max drawdown and fold-1 instability prevent a higher score. - **Fundamentals trend (90/100):** NVDA and TSM both post top-decile revenue growth, expanding margins, and strengthening balance sheets — the fundamental case is nearly unimpeachable.

Trade now

**Status: WAIT — no entry conditions are live across any of the three target tickers.** The strategy fires when a semiconductor name gets hammered on guilt-by-association while the broader market stays green. The specific trigger requires price trading below the lower Bollinger Band (20, 2σ), an RSI reading below 35, and an ADX above 20 — all on the daily chart. Right now, none of the three targets are in the entry zone. **NVDA at $207.40** is the closest to actionable but still short. The Bollinger lower band sits at $201.78 — NVDA needs to drop roughly $5.62 (about 2.7%) to reach that band. RSI (17) is at 55.4, well above the 35 trigger; it would need to fall 20 points. ADX at 68.8 is met, confirming trend strength is present. The nearest support level for NVDA is $208.82, and the strategy also wants price to probe that support and reclaim it — possible but not yet in evidence. **SMH at $566.93** is already stretched: price is below the Bollinger band ($596.76), RSI has crashed to 12.2, and ADX is maxed at 100. But SMH is an ETF, not a single stock — the thesis centers on buying individual semiconductor names caught in the crossfire, and the strategy's position sizing and support-level rules are calibrated per-ticker. **TSM at $409.74** is also near trigger: price is below its Bollinger band ($432.28), RSI is at 35.7 (just 0.7 points above the 35 threshold), but ADX at 12.4 is well below the 20 requirement — the strategy needs a stronger trend signal before it would consider TSM. **If entering on trigger**, the stop loss is set at 2.5% below entry and the take-profit at 4.9% above, yielding an effective reward-to-risk of roughly 2:1. The strategy also enforces a 45-bar time stop. The walk-forward holdout using the recommended RSI period 17 produced a 43.3% return across 28 trades with a 60.7% win rate and a 10.4% max drawdown, but that record was built on setups where all conditions aligned — it is not a reason to enter before they do. "Wait" means setting alerts at the Bollinger lower band levels and monitoring for intraday divergence between SPY (green) and the target ticker (down 3%+).

Why the dip-buying thesis has real support

The idea argues that the IBM-driven tech panic is creating a buying opportunity in fundamentally strong semiconductor stocks that are simply caught in the crossfire. The numbers strongly support the quality-of-underlying-assets side of that thesis. NVIDIA's FY2026 results show revenue of $215.9B with 65.5% year-over-year growth, a gross margin of 71.1%, and an operating margin of 60.4% — placing it in the 99th percentile…

Scores

• Conviction score breakdown: 63
• Thesis support: 72
• Trade readiness: 38
• Risk quality: 45
• Backtest evidence: 68
• Fundamentals trend: 90

Watch items

• NVDA — RSI (17)
• NVDA — Price vs Bollinger Lower Band (20, 2σ)
• TSM — RSI (17)
• TSM — ADX (14)
• SMH — RSI (14)
• NVDA — Next earnings report
• NVDA — Next earnings report
• NVDA — Price below Bollinger (20)
• NVDA — RSI (17) below 35
• NVDA — ADX (14) above 20