While Tesla's earnings miss has dragged down tech headlines, the underlying demand for AI computer chips is actually accelerating. Alphabet is leading a new wave of spending among the mega-cap tech companies, and Nvidia's monumental half-trillion-dollar d

Verdict: compelling thesis, unproven strategy — wait for triggers

The AI capex thesis is real: Nvidia's 71.1% gross margin, 99th-percentile operating margin, and $96.8B in free cash flow confirm the build-out is translating into extraordinary profitability, and AMD's 34.3% revenue growth with a fortress balance sheet (debt-to-equity of 0.037) provides a credible second leg. The trade killer is statistical fragility — the 210.3% backtest return comes from just two AMD trades over 60 months, the walk-forward was rejected for insufficient samples, no robust parameter setup was established, and NVDA generated zero triggers in either evaluation window. Both tickers currently sit well below their entry gates, with RSI far below price and AMD's ADX at 18.1 still short of the 20 threshold, making this a setup waiting for conditions rather than an actionable signal today. **Conviction Breakdown** - **Thesis support (80):** The fundamental case is genuinely strong — per the cited Reuters and CNBC reports, the capex cycle is accelerating, and both companies' financials confirm monetization. - **Trade readiness (15):** Multiple entry conditions are unmet on both tickers; RSI remains far below price and the strategy is not actionable at current levels. - **Risk quality (35):** The 23.8% max drawdown is meaningful, exits were simulated on daily bars rather than intraday data, and the near-zero historical correlation between AMD and NVDA could flip positive in a sector-wide selloff. - **Backtest evidence (20):** Two trades cannot distinguish skill from luck; walk-forward validation failed the six-trade minimum, and the 12-month holdout produced zero triggers. - **Fundamentals trend (85):** Both tickers show accelerating revenue, improving margins (AMD's Q3 gross margin recovered to 51.7% after a Q2 dip to 39.8%), and minimal leverage constraints on continued investment.

Trade now

**Action today: Wait.** The strategy needs price at or below the 50-day moving average, RSI (14) above price, and ADX (14) above 20 — plus a touch of the nearest support level. Price is currently below that moving average on both tickers, which is met, but RSI above price is the hard blocker. AMD's RSI sits at 38.5 against a closing price of $494.95 — RSI is well below price, so that condition is not there yet. NVDA is closer in spirit at an RSI of 33.6, but with price at $196.51 the same gap applies. ADX is mixed: AMD at 18.1 needs to climb above 20 (short by roughly 2 points), while NVDA at 21.9 has already cleared that bar. **Where you stand on exits.** The take-profit is set at 4.9% above entry and the stop at 2.4% below, giving an effective reward-to-risk of roughly 2:1. Fibonacci-based exits also apply: a 127.2% extension for take-profit and a 61.8% retracement for stop-loss. On the overbought-signal exit, RSI would need to exceed 75 — currently at 33.6 on NVDA and 38.5 on AMD, so nowhere near triggering. The holding-period exit (60 bars) and EMA(9) cross only activate once a position is open. **What "wait" means concretely.** Set alerts for AMD's RSI (14) crossing above its closing price and for ADX crossing above 20 — those are the two missing gates. For NVDA, the only remaining gate is RSI crossing above price. Until one of those fires, no entry signal is valid. No robust parameter setup was established in walk-forward testing (fewer than 6 trades across folds), so the configured baseline rules stand as-is. **Backtest context.** The AMD leg of this strategy returned 210% over 60 months across 2 trades with a 50% win rate and a 23.8% max drawdown. The NVDA leg produced no triggers in a 24-month evaluation window. Exit fills were simulated on daily bars, not intraday, so reported drawdown and win rate should be treated as approximate.

Why the AI chip demand thesis has teeth

The thesis rests on an accelerating AI capex cycle, and the cited news flow gives it real weight. Per the Reuters report from July 24, Nvidia and SK Group unveiled a data center initiative exceeding $500 billion, alongside a memory partnership that underscores the sheer scale of the build-out. The CNBC piece the same day notes that Alphabet is raising the stakes on AI spending among hyperscalers, with three more mega-cap platforms yet to report. The idea argues this spending wave is…

Scores

• Conviction score breakdown: 47
• Thesis support: 80
• Trade readiness: 15
• Risk quality: 35
• Backtest evidence: 20
• Fundamentals trend: 85

Watch items

• AMD — RSI (14)
• AMD — ADX (14)
• NVDA — RSI (14)
• NVDA — RSI (14)
• AMD — Price vs EMA (9)
• NVDA — ADX (14)
• AMD — Price
• AMD — RSI (14) above Price
• AMD — ADX (14) above 20