Every AI data center needs enormous amounts of the same memory chips that go into iPhones, so Nvidia's forecast of AI demand extending into 2028 effectively guarantees a years-long memory supply squeeze. Tim Cook has already flagged this as a once-in-a-ce
Every AI data center needs enormous amounts of the same memory chips that go into iPhones, so Nvidia's forecast of AI demand extending into 2028 effectively guarantees a years-long memory supply squeeze. Tim Cook has already flagged this as a once-in-a-century price shock, which means Apple either absorbs rising component costs (crushing its famous hardware margins) or raises prices and risks selling fewer devices. Hardware companies like Apple are the price-takers in this squeeze, while the chip companies driving demand are the beneficiaries. A long-NVDA/short-AAPL pair captures that transfer of profit from device makers to AI infrastructure.
Idea
Every AI data center needs enormous amounts of the same memory chips that go into iPhones, so Nvidia's forecast of AI demand extending into 2028 effectively guarantees a years-long memory supply squeeze. Tim Cook has already flagged this as a once-in-a-century price shock, which means Apple either absorbs rising component costs (crushing its famous hardware margins) or raises prices and risks selling fewer devices. Hardware companies like Apple are the price-takers in this squeeze, while the chip companies driving demand are the beneficiaries. A long-NVDA/short-AAPL pair captures that transfer of profit from device makers to AI infrastructure.
Advanced Analysis — institutional-depth research report
Verdict: the memory squeeze is a real story, but the numbers say wait
**Verdict: interesting thesis, but the numbers say wait.** The idea's strongest point is the stark fundamental gap: Nvidia grew revenue 65.5% to $215.9B with a 71.1% gross margin while Apple grew just 6.4% to $416.2B, and Nvidia's AI demand forecast extending into 2028 (per the Yahoo Finance coverage) is exactly the demand shock a memory squeeze thesis needs. Its biggest weakness is that Apple's own numbers contradict it — quarterly gross margin has climbed from 46.2% to 50.1% across the five most recent quarters even after Cook's "100-year flood" warning, and the backtest itself endured a mid-2022 stretch down roughly 13% with a 48.8% win rate. Meanwhile, no robust parameter setup was established for the entry rules, and none of the live entry conditions are triggered: NVDA's RSI (14) sits at 50.9 versus the required at-or-below-45, AAPL's at 56.2, and both ADX readings (3.2 and 1.0) are far below the 25 trend-strength floor. What would flip the verdict: Apple's next earnings call showing gross margin compression from memory costs, alongside NVDA's price breaking below its 21-day EMA with RSI at or below 45. Until the memory squeeze shows up in Apple's reported margins and the entry rules fire, patience is the trade. **Conviction breakdown** - Thesis support: 55 — pricing-power narrative is coherent and the demand data supports it. - Trade readiness: 25 — every entry condition remains unmet on both legs. - Risk quality: 45 — near-zero pairwise correlation (-0.093) helps, but a two-name basket rests on one correlation estimate. - Backtest evidence: 60 — 34.5% return over 60 months and 80 trades, but coarse daily fills and a tough 2022 stretch temper it. - Fundamentals trend: 35 — Nvidia's numbers are superb; Apple's rising margins argue against the short leg.
Trade now: armed but waiting on momentum and trend filters
## Trade now: wait — the entry rules are close but not triggered NVDA closed at $217.55, about 0.8% above its 21-day EMA of $215.71 — the closest entry condition to firing, but not there yet. AAPL closed at $313.77 versus its 21-day EMA of $311.98, roughly 0.6% above. Both names need price below that EMA to start an entry, and neither qualifies today. The bigger gaps are in momentum and trend strength. NVDA's RSI (14) sits at 50.9 — the rules need it at or below 45, about 5.9 points away. AAPL's RSI is 56.2, needing roughly 11 points of cooling. Both tickers' ADX (14) readings are far from the required trend-strength condition: NVDA at 3.2 and AAPL at 1.0, versus a requirement of 25 or above (a 20 floor also applies). ADX that low means both stocks are choppy and trendless — exactly the environment this setup is designed to avoid. What waiting means in practice: no orders until each leg's entry conditions are live — price below the 21-day EMA, RSI (14) at or below 45, and ADX (14) at or above 25. If triggered, the strategy's fixed-risk sizing uses a 2.6% stop and a 5.1% profit target, roughly a 2.0 reward-to-risk ratio per position, with exits also managed at the second resistance level and below first support (NVDA's first support is $210; AAPL's is $310). The completed backtest supports patience: 34.5% total return over 60 months across 80 trades, a 48.75% win rate, and a 15.7% maximum drawdown — entries taken with the ADX filter intact have historically been the higher-quality subset. One caveat: exits in that backtest were filled on daily trigger bars, so the reported drawdown and win rate are coarse rather than precise to intraday moves.
unused
The core of this idea — long NVDA, short AAPL, on the theory that AI demand transfers pricing power from device makers to AI infrastructure — starts with a stark fundamental gap. Nvidia's latest fiscal year shows revenue of $215.9B, up 65.5% year over year, with a 71.1% gross margin, a 60.4% operating margin, and $96.7B of free cash flow. Apple, by contrast, grew revenue just 6.4% to $416.2B on a 46.9% gross…
Scores
- Conviction score breakdown: 44
- Thesis support: 55
- Trade readiness: 25
- Risk quality: 45
- Backtest evidence: 60
- Fundamentals trend: 35
Watch items
- NVDA — RSI (14)
- NVDA — ADX (14)
- NVDA — Price vs 21-day EMA
- NVDA — First support level
- AAPL — RSI (14)
- AAPL — ADX (14)
- AAPL — Nearest resistance
- AAPL — Price below EMA (21)
- AAPL — ADX (14) above 25
- AAPL — RSI (14) below 45
- AAPL — ADX (14) above 20