TSMC's record August sales — up more than half versus a year ago — are hard evidence that AI chip demand is real money in the till, not just hype, and reporting on a coming memory shortage suggests chipmakers can keep raising prices. Unlike crypto or rate

Verdict: a strong thesis that hasn't earned your money yet

The verdict: wait — the thesis is strong, but the trade is not armed. The single strongest point for the idea is that TSMC's demand story is confirmed in hard numbers: fiscal 2024 revenue grew 33.9% to NT$2.89 trillion, free cash flow jumped about 198%, and the August revenue report per CNBC (September 10, 2026) showed a record surge of over 53%. The strongest point against is that both entry legs are far from firing — QQQ's 14-day RSI at 51.9 and TSM's at 65.5 need to fall to 45 and then cross back up through 40 — while the backtest edge (96.2% over 60 months on just 8 trades at a 25% win rate) depends entirely on entering only on pullbacks, with the 12-month window unevaluable for lack of history. The QQQ hedge is also softer than it looks: QQQ is 60.9% technology with a 44.9% top-10 weight of TSMC's own customers, and the covered look-through revenue growth on that weight is negative 10.1%, so a macro shock hits both legs together. Insider buying of about $370,000 across 20 holders as of the June 30, 2026 reporting period is mildly supportive, though that filing cycle is already closed and later activity is invisible. What would flip the verdict: a pullback that drags QQQ RSI to or below 45 and then turns it back up through 40 while price holds above the 50-day EMA at $710.57 — that is when this becomes a buy.

Trade now: the setup is live on trend, but not yet on pullback — wait for the RSI window

Both entry legs are trend-confirmed today but fail the pullback test. On QQQ at $716.31, price is above its 50-day EMA ($710.57) and the ADX (10.3) is below 25 — both conditions met — but the 14-day RSI reads 51.9, well above the required 45, and the rule also needs RSI to have crossed up through 40, which is roughly 12 points of RSI away. On TSM at $435.56, price is $16.53 above its 50-day EMA ($419.03), but RSI is 65.5 — more than 20 points above the 45 cap — and ADX (27.3) is also slightly above the 25 ceiling. In plain terms: both stocks are too strong right now for this strategy's dip-buy entry, and the concrete meaning of "wait" is a pullback that drags RSI down through 45 and then turns it back up through 40 while the trend stays intact. The strategy's protective machinery, once a position exists, is a stop at a 2.3% loss against a 4.6% take-profit, an approximate 2:1 reward-to-risk per trade, plus an exit if the 50-day EMA crosses below the 200-day EMA or RSI breaks 35. On QQQ, the 50-day EMA would need to fall about 42 points to reach the 200-day EMA ($668.36); on TSM about 39 points ($379.99) — both far from today's readings, so the thesis invalidation is distant. The 60-month backtest on the QQQ leg returned 96.2% across 8 trades with a 25% win rate and a 10.3% maximum drawdown, so the edge came from winners far larger than losers, which is exactly why you should not chase strength outside the rules: the entry discipline is where the risk control lives. Note that no robust alternative parameter setup was established — the parameter study could not be evaluated on available history, so the published trigger levels above are the ones to follow, unmodified. If you want exposure before the triggers fire, size it outside this system and treat it as a separate decision; the strategy's own answer today is patience.

Record August Revenue and a 96% Backtest: The Demand Story Is Real

The idea's core claim — that AI chip demand is real money, not hype — is well supported by the fundamentals. In fiscal 2024 TSMC grew revenue 33.9% to $2.89T (NTD-basis, per SEC normalized facts), lifted net income 36.0% to $1.16T, and expanded free cash flow 197.9% to $870.2B. That free cash flow sits in the 99.9th percentile among 791 Information Technology peers, and the 45.7% operating margin lands in the 98.4th percentile of 854 peers. This is not a story stock; the cash is already in the till. The news flow lines up with these numbers: per the CNBC piece from September 10, 2026, August revenue surged over 53% to a record high, and the Yahoo Finance article on the AI memory crunch describes pricing conditions approaching '10 out of 10' craziness — the setup the idea argues lets chipmakers keep raising prices. The business economics back the pricing-power argument. Gross margin recovered from 54.4% in 2023 to 56.1% in 2024, net margin held above 40%, and return on equity improved from 24.8% to 27.3%. Meanwhile the balance sheet remains almost comically unlevered: debt-to-equity, even after a 486.7% jump, is still just 0.75%. With $2.13T of cash on hand against $31.8B of long-term debt, the 'fundamental cushion' the idea leans on is genuine. The dividend record adds a capital-return kicker: trailing payments of $3.57 per share over the last twelve months versus $2.71 the prior year, with annual dividends up 31.9%. On the strategy evidence itself, this is a backtested setup, and the completed backtest supports the thesis direction. Over the 60-month window on daily bars, the rule set produced a 96.2% total return across 8 trades with a maximum drawdown of…

Scores

• Conviction score breakdown: 58
• Thesis support: 85
• Trade readiness: 30
• Risk quality: 45
• Backtest evidence: 45
• Fundamentals trend: 85

Watch items

• QQQ — RSI (14)
• QQQ — Price vs 50-day EMA
• QQQ — ADX (14)
• TSM — RSI (14)
• TSM — ADX (14)
• TSM — Price vs 50-day EMA
• QQQ — RSI (14)
• TSM — 50-day EMA vs 200-day EMA
• TSM — Insider ownership filings