AI spending fears are going global — Asian chips and copper break down, short the US chip basket

Verdict: the AI contagion case is live, but the trade is not armed — wait for confirmation

The verdict on this AI-de-rating short is: wait. The strongest point for the trade is the breadth of the contagion — per CNBC, Reuters, and Bloomberg on October 13, 2026, Asian chip stocks, US sentiment, and even copper and aluminum all cracked together, which is exactly the mechanism the thesis needs. The strongest point against is that the thesis is fighting two of the strongest income statements in the sector: NVDA's July-quarter revenue rose 17.9% sequentially to $96.2B with a 66.2% operating margin, and AMD's revenue grew 12.5% to $11.5B with net income up 66%. The cash-flow lines are the bear case's best ammunition — NVDA's free cash flow fell 56% to $21.4B, its debt-to-equity jumped to 0.146, and the provisional September 30 filings show roughly $569.5M net insider selling at NVDA and $176.2M at AMD — but these are still provisional and recent, not confirmed trends. Compounding the caution, the mechanical entry rules produced zero entries across 1,235 daily bars over 60 months, and parameter sensitivity returned no robust setup, so the watch levels are published as-is rather than validated. A follow-up NVDA print showing free cash flow re-accelerating would flip the verdict against the short entirely.

Trade now: the basket short is still a watch-list, not a fill

Nothing in this setup is live yet, and that is the honest headline. The entry rules require, on each of SMH, NVDA and AMD, a close that has crossed above its 20-day average while trading below it, trend strength (ADX, 14) above 20, and RSI (14) below 50. Today SMH closes at $590.99 versus its 20-day average of $601.08 — below it, but no fresh cross — with RSI at 36.1 and ADX at 17.5, still under the 20 threshold. NVDA is closer: $226.27 against a $228.47 average, RSI 41.3, ADX 16.9. On the supplied feed, AMD shows all conditions met except the fresh cross, closing $594.24 against $597.03 with RSI 41.3 and ADX 25.2. So 'wait' means one thing concretely: do not pre-position. The conditions are evaluated on daily closes, and the nearest gaps are small — SMH needs a new close back above its 20-day average plus an ADX lift of about 2.5 points to 20; NVDA needs the same cross, roughly $2.20 above Tuesday's close, plus an ADX lift of about 3.1 points. Until a full daily close satisfies every condition on a symbol, there is no position to size. One tension you should see plainly: the idea argues for riding AI contagion lower, yet the compiled entry rules are written as long entries on the same names (price crossing above its 20-day average while below it, with RSI under 50 — a reversion-style condition). We read the research intent as the bearish de-rating case and the rules as the mechanical trigger layer, and we flag the mismatch rather than paper over it. Note also that no robust parameter setup was established — the sensitivity search returned no recommendation — so the thresholds above are used as published, not optimized. If an entry does trigger, the published exits are mechanical: a hard stop at a 2.5% loss on the position, a take-profit at 4.9%, an exit if the trade is held 30 daily bars, and level-based exits at the first support/resistance levels (for example, SMH's nearest support at $590 and nearest resistance at $595.48). Those percentages define the effective reward-to-risk at roughly 2:1 on any filled position. The rules were evaluated on real daily bars over the past five years (1,235 bars) but did not open an entry — this is a watch-list setup, not an active signal.

The Bull Case: Earnings Firepower Behind the Chip Basket

If you are considering the bearish thesis on the semiconductor basket, you have to first size up what you are fighting. NVIDIA's most recent quarter (period ending July 26, 2026) delivered $96.2B in revenue, up 17.9% quarter over quarter, with a 66.2% operating margin and $59.7B of net income. Those are not the numbers of a trade that is rolling over — they sit at the 99th percentile of operating margin and the 92nd percentile of revenue growth among 899 Information Technology peers. The idea's contagion narrative is fighting the strongest income statement in the sector. AMD is also accelerating, not decaying. Its June 27, 2026 quarter showed revenue of $11.5B, up 12.5% sequentially; net income jumped 66.1% to $2.3B; net margin expanded from 13.5% to 19.9%; and gross margin ticked up to 53.8%. AMD's trailing fiscal-year revenue growth ran at 34.1%, placing it in the 85th percentile of its sector. Both companies are compounding, which historically makes broad de-rating shorts expensive to hold for long. The balance sheets give the bulls staying power too. AMD's debt-to-equity actually declined quarter over quarter to 0.048, and NVDA's sits at just 0.146 even after a 236% increase. Neither name needs the market to fund its growth, so a risk-off week cannot create financing pressure the way it would for a levered issuer. NVDA also reduced shares outstanding slightly, from 24.2B to 24.1B. Note the evidence tier honestly: the entry rules evaluated on real daily bars did not fire at all in 1,235 bars over 60 months (nor in 24- and 12-month windows), so this is a watch-list setup,…

Scores

• Conviction score breakdown: 42
• Thesis support: 60
• Trade readiness: 20
• Risk quality: 45
• Trigger proximity: 30
• Fundamentals trend: 55

Watch items

• SMH — Close vs EMA (20)
• SMH — ADX (14)
• SMH — RSI (14)
• NVDA — Close vs EMA (20)
• NVDA — ADX (14)
• AMD — Close vs EMA (20)
• AMD — Insider open-market net flow
• NVDA — Insider open-market net flow
• NVDA — Free cash flow
• AMD — Net income