Building 20 gigawatts of AI computing capacity is a staggering amount of hardware, and every one of those servers needs large quantities of memory and storage chips. The companies that make them — Micron, Western Digital, and Seagate — are the direct supp

Verdict: real cycle, wrong moment — wait for the trigger

This is one of the rare ideas where the fundamental inflection is already in the numbers, not just in the narrative: per the supplied data, Micron's quarterly gross margin ran from 39.8% to roughly 56%, 74.4%, then 84.6% in successive quarters, and free cash flow swung from -$6.1B in fiscal 2023 to quarterly prints of $3.0B, $8.5B, and $26.1B — exactly the pricing-power signature the 20-gigawatt AI buildout thesis (per the Yahoo Finance piece) predicts. The strongest argument against is that memory is a brutal commodity cycle and today's 84.6% margin is precisely the kind of reading that historically invites supply and collapse — the same companies posted negative margins just two years ago — while the trade statistics are far weaker than the headline: the 241.5% five-year return came from a single trade, and the recent 24-month window produced a 20% win rate across 5 trades with roughly 30% drawdowns in both windows. Execution-wise, none of the three names is at a fresh entry today, and the basket's pairwise correlations of 0.63–0.84 mean you are buying one memory-cycle bet three times, not a diversified trio. **Conviction breakdown:** Thesis support 72 (fundamentals inflecting before the news broke, per the idea's own evidence); Trade readiness 35 (no entry conditions met; STX blocked by ADX at 1.5, WDC needs a 10.7% rally); Risk quality 45 (2.6% stop is very tight for stocks this volatile, and correlations drift toward 1 in drawdowns); Backtest evidence 40 (single 60-month trade; 20% win rate recently; exits filled on daily bars); Fundamentals trend 85 (top-3% sector free cash flow across all three names, margins still rising).

Trade now

**Where things stand.** The strategy goes long each name when price crosses back above its 50-day average, RSI (14) crosses above 45, and ADX (14) is above 20. None of the three is in a fresh entry today — the crossover conditions need a reclaim that hasn't happened yet for STX or WDC, and MU is already well extended above its trigger. **Name by name.** MU closed at $966.78, roughly $60 above its 50-day average of $907.14 — the cross already happened long ago, so this is not a fresh entry; it's mid-trend, with RSI at 57.5 (12.5 points above the 45 cross level) and ADX at 32.0, comfortably above 20. STX is the near-miss: it closed at $850, just $1.83 below its 50-day average of $851.83, RSI at 45.5 (barely above 45), but ADX at 1.5 — a low directional-strength reading that is 18.5 points short of the 20 requirement, so the entry is blocked by trend strength, not price. WDC closed at $459.44, about $49 below its 50-day average of $508.32 and 3.4 points below the RSI 45 threshold, though its ADX of 23.2 already clears the bar. **Risk framing if an entry fires.** The strategy takes profit at +5.3% and stops out at -2.6%, an effective reward-to-risk of roughly 2:1 per trade, with a Fibonacci 78.6% stop and 127.2% target layered on top. "Wait" here means something concrete: don't buy strength in MU mid-trend; watch STX for a close above $851.83 with ADX rising through 20, and WDC for a close above $508.32 with RSI crossing 45. Until those prints occur, there is no trade under the rules. **Evidence base.** The backtest sample is thin — one completed MU trade over 60 months returning 241.5%, and a 24-month window with 5 trades at a 20% win rate and a peak drawdown of about 31% (roughly 30% in the longer window too). Treat the headline return as one large winner, not a repeatable hit rate, and size accordingly: the fixed-risk sizing caps each position at 2.65% risk of equity, with positions no larger than 25% of the book.

The order books already back the thesis

The bull case here is not speculative — it is already showing up in the income statements. Micron's latest full year showed revenue of $37.4B, up 115.3% year over year, a growth rate that puts it in the 71st percentile of the Information Technology sector. More striking is the trajectory in the quarterly series: gross margin went from 39.8% at fiscal year-end August 2025 to roughly 56% in the November quarter, then 74.4%, then 84.6% by May 2026. That is the signature of an industry where demand is outrunning supply — exactly what the idea argues a 20-gigawatt AI buildout would do to memory pricing. The Yahoo Finance piece on Musk's 20 gigawatts of compute capacity is the…

Scores

• Conviction score breakdown: 55
• Thesis support: 72
• Trade readiness: 35
• Risk quality: 45
• Backtest evidence: 40
• Fundamentals trend: 85

Watch items

• STX — ADX (14)
• STX — Close vs EMA (50)
• WDC — Close vs EMA (50)
• WDC — RSI (14)
• MU — RSI (14)
• MU — Close vs nearest support
• MU — Price crossed above EMA (50)
• MU — Price
• MU — RSI (14) crossed above 45
• MU — ADX (14) above 20