Big Tech spending panic meets booming AI demand — long the AI chip suppliers

Verdict: The thesis is real, but the setup isn't armed yet — wait for the pullback

Micron's quarter ended May 28, 2026 is the single strongest point in this idea's favor: revenue of $41.5B, up 73.8% sequentially, a gross margin of 84.5%, free cash flow of $26.1B, and debt-to-equity down to 5.1% — filings-level proof the AI memory cycle is running hot, and per Barron's on June 25 it revived the AI trade. The strongest point against is uncomfortable ownership evidence: filings for the period ended June 30, 2026 show net open-market insider selling of roughly $231M at Micron and $375M at NVIDIA, plus Micron's own fiscal-2023 history of gross margins collapsing to negative 32.7% as a reminder of how fast memory economics can mean-revert. On execution, nothing is live: the strategy's RSI-at-or-below-40 entry gate sits roughly 18.8 points away on NVIDIA (RSI at 58.8), so the trade we would otherwise take has not armed, and the completed five-year backtest (52 trades, 53.8% win rate, 10.6% maximum drawdown, 10.8% total return) describes a setup that profits from buying pullbacks — and there is no pullback yet. A fresh ownership filing for the June 30 period, or Micron's next quarterly report around early October, would materially change the read. For now, the verdict is wait: set alerts on the $220.54 first support and an RSI reading of 40, and let the tape come to you.

Trade now: the entry gate is wide of the mark — stay flat and set the alert

Nothing is live today. The strategy's long entry requires the trade symbol to sit at or above its first support level, with its 14-day RSI at or below 40, a one-day rate of change above 0.5, and an active 10-day channel. On NVDA, the intended trade vehicle, three of those four conditions are already met — but the RSI reads 58.8 against the 40 threshold, roughly 18.8 points away. The stock closed at $224.41, so the RSI pullback is the only gate still open. The Microsoft-based variant is in the same position; both entries are waiting, not broken. If an entry triggers, the risk plan is mechanical: a 2.4% fixed stop against a 4.9% take-profit, with a wider 15% profit exit available. That is roughly a 2-to-1 reward-to-risk ratio per position, sized so no single name exceeds 25% of the book. "Waiting" therefore means setting alerts on NVDA at the $220.54 first support and on an RSI reading falling to 40 — not chasing a stock that just closed above its 50-day average of $209.45. Patience is supported by the completed backtest on the NVDA pair: over the five-year window it produced 52 trades at a 53.8% win rate, a 10.8% total return, and a 10.6% maximum drawdown, with the shorter 12-month window showing 7 trades at a 71.4% win rate. The edge comes from buying pullbacks, and there is no pullback yet. One planning note: the parameter-sensitivity evaluation ran out of its time budget, so no nearby-parameter recommendation was established — trade the published rules as written.

Micron's numbers say the AI build-out is real — and the pullback may be the entry

The thesis — long the AI hardware suppliers while big-brand tech gets repriced — is anchored by Micron's own filings. For the quarter ended May 28, 2026, Micron reported revenue of $41.5B, up 73.8% from the prior quarter's $23.9B, with net income of $28.2B and a gross margin that climbed to 84.6% from 74.4% one quarter earlier. That is not a story stock; it is a balance-sheet event. Free cash flow jumped to $26.1B and the debt-to-equity ratio fell to 0.051 from 0.132, meaning Micron is funding the build-out largely from its own cash generation. Per Barron's June 25 coverage, those earnings revived the AI trade and lifted Nasdaq futures. The same inflection shows up across the reported series: Micron's operating margin went from 26.1% in the fiscal year ended August 2025 to 80.4% in the quarter ended May 2026, and return on equity reached 28.0%. The suppliers thesis also has legs in TSMC, which grew revenue 33.9% in 2024 to $2.89T NT ($2.89T local currency) with a 40.0% net margin, and NVIDIA, whose quarter ended April 26, 2026 posted $81.6B of revenue (up 19.8% sequentially) with a 71.5% net margin. These are the companies actually selling the physical AI infrastructure, exactly as the idea argues. On execution, the completed backtest gives the rule set a concrete record: over the 60-month daily window the strategy completed 52 trades with a 10.8% total return, a 53.8% win rate, and a worst peak-to-trough drawdown of 10.6%. On the tighter 24-month window it delivered a 7.95% return across 18 trades at a 61.1% win rate, and over the last 12 months a 2.1% return on just 7 trades with a 71.4% win rate. The win rate improving as the window shortens is consistent with the idea that the divergence trade (buying suppliers when big tech sells off) has worked better recently. Finally, the setup's trigger logic fits the cited news almost perfectly: per MarketWatch, Microsoft suffered a historic June rout as investors balked at heavy capital spending, and per CNBC, Apple was slammed on Mac and iPad price hikes. A rule that buys the supplier when the buyer's stock drops more than 1% is literally constructed to harvest those exact sessions. No robust…

Scores

• Conviction score breakdown: 59
• Thesis support: 78
• Trade readiness: 25
• Risk quality: 55
• Backtest evidence: 65
• Fundamentals trend: 72

Watch items

• NVDA — RSI (14)
• NVDA — Close vs first support
• NVDA — Net open-market insider value
• MU — Quarterly revenue trajectory
• MU — Net open-market insider value
• TSM — Net open-market insider value
• MU — RSI (14) below 40
• MU — ROC (1) above 0.5
• MU — Donchian (10) above 1
• NVDA — RSI (14) below 40
• NVDA — ROC (1) above 0.5