AI hardware demand is printing real cash — long Micron and the chip leaders

Verdict: real cash, extreme cycle — wait for the trigger

This idea's core claim — that AI demand is showing up as real cash, not just guidance — is hard to dispute: Micron's fiscal Q3 (ended May 28, 2026) delivered $41.5B revenue, up 73.8% sequentially, free cash flow of $17.6B, and an 84.6% gross margin. But the strongest point against is that the fundamentals sit at a cyclical extreme: Micron's gross margin was negative just three years ago, the annual dividend actually fell 27.8% in 2026, and provisional ownership filings for the period ended September 30, 2026 show roughly $231M of net insider selling at Micron and $569M at Nvidia. The trade itself is also armed but not triggered — price sits above the EMA (21) and Bollinger (20) with RSI at 58, but the stochastic crossover (fast line at 79.4 versus slow at 55.6) has not printed. The completed backtest supports the rule set (74.1% return over five years, 8.8% max drawdown, but a 46.7% win rate across 45 trades), while no robust nearby-parameter setup was established because the sensitivity evaluation ran out of its time budget. What flips the verdict: a confirmed daily stochastic cross above the slow line while price holds above the EMA (21), or a mid-October ownership filing showing the insider selling reversed. Until then, wait — the idea is good, the entry is not yet live.

Trade now

**Status: armed but not triggered.** MU closed at $1,064.54, and three of the four entry conditions are already live: price is above its Bollinger (20) level of $1,026.88, above the EMA (21) at $1,031.26, and the RSI (14) sits at 58.0, comfortably above the required 40 threshold. The one missing piece is the momentum trigger — the fast Stochastic (14) at 79.4 has not crossed above the slower Stochastic (3) at 55.6. Until that crossover prints on the daily chart, there is no entry. "Wait" here means literally sitting on your hands: a pullback followed by a fresh stochastic cross, or a strong up day that closes with the fast line overtaking the slow one, is what flips this on. If triggered, the risk framework is mechanical. Position sizing uses fixed risk at roughly 2.5% of the account per trade, capped at 25% of capital per position, with a hard stop at a 2.5% loss and a take-profit at a 4.9% gain — an effective reward-to-risk of about 1.9-to-1. Secondary exits apply as well: an overbought RSI exit above 75, a take-profit at the 127.2% Fibonacci extension, and stops below the second support level or the 61.8% retracement. For MU, the nearest support levels sit at $1,011.77 and $1,000.00, which is where the structural stop would tighten. The evidence base is a completed backtest on the same rules: over the five-year window on MU the setup produced 45 trades, a 46.7% win rate, a 74.1% cumulative return, and a maximum drawdown of 8.8%. The shorter twelve-month window ran 13 trades at a 38.5% win rate with a 16.2% return — a reminder that this system is streaky and relies on letting the roughly 2-to-1 payoff do the work, not on high hit rates. One caveat on parameters: the sensitivity evaluation exceeded its time budget, so no robust nearby-parameter setup was established; trade the rules as written. Do not chase the trade early — entering before the stochastic cross changes the odds you're taking.

Demand, Cash, and the Tape Are All Pointing the Same Way

The idea's core claim — that AI demand is now showing up as real orders and real cash, not just guidance — is well supported by Micron's latest quarter (fiscal Q3, ended May 28, 2026). Revenue jumped 73.8% quarter over quarter to $41.5B, gross margin expanded roughly ten points to 84.6%, and free cash flow more than tripled to $17.6B from $5.5B the prior quarter. Net income more than doubled to $28.2B with a 68.1% net margin. Those are the numbers of a company at a cyclical peak of pricing power, which is exactly the fuel the thesis needs for the 'laggard catches up' argument. The cash-conversion leg of the thesis also checks out against peers. In its most recent annual period (ended August 2025), Micron posted 46% year-over-year revenue growth, ranking in the 89th percentile of 897 Information Technology peers, with a 96th-percentile operating margin. NVIDIA, the demand engine behind the whole trade, reported $96.2B of quarterly revenue (up 17.9% sequentially) in the quarter ended July 26, 2026 with a 75% gross margin — confirming the AI build-out is still accelerating rather than decelerating. Per the Reuters piece from October 5, 2026, Foxconn's blowout third quarter beat forecasts on AI demand, reinforcing that hardware orders are flowing through the chain now, not at some vague future date. The rule-based strategy tied to this idea has a completed, traded backtest behind it. Over the full 60-month daily window the MU-focused setup produced a 74.1% return across 45 trades, with a maximum drawdown of 8.8% — a favorable return-to-drawdown profile. The shorter windows corroborate rather than contradict: the 24-month window returned 36.1% with a 7.6% max drawdown, and the most recent 12 months returned 16.2% with a 9.1% drawdown. Positive returns in all three windows, including the most recent one, mean the momentum thesis has been earning its keep recently, not just in the distant past. Finally, the speculative…

Scores

• Conviction score breakdown: 62
• Thesis support: 78
• Trade readiness: 45
• Risk quality: 55
• Backtest evidence: 62
• Fundamentals trend: 70

Watch items

• MU — Stochastic (14) vs Stochastic (3) crossover on MU daily
• MU — MU price vs EMA (21)
• MU — MU RSI (14) vs 75 exit threshold
• MU — MU price vs second-ranked support
• NVDA — NVDA RSI (14) vs 75
• SMH — SMH RSI (14) vs 75 exit threshold
• MU — Insider net open-market activity (period ending 2026-09-30)
• MU — MU next quarterly earnings (fiscal Q4)