$35 billion in Nvidia orders from China's tech giants while AI stocks are beaten down — buy the dip

Verdict: the demand story is real, but the entry hasn't fired — wait

The idea's core demand claim is genuinely testable and currently supported: the Reuters report of over $35B in Nvidia chip orders from ByteDance, Alibaba, and Tencent, combined with NVDA's fiscal Q4 showing 73.2% year-over-year revenue growth and an operating margin in the 99.3rd percentile among IT peers, makes this one of the few AI-dip theses backed by checkable fundamentals. The strongest point against it is the freshest data: NVDA's fiscal Q2 (ended 2026-07-26) shows free cash flow down 56% sequentially to $21.4B, net margin compressed 9.4 points to 62.0%, and the provisional 2026-09-30 ownership filings show roughly $569M of net insider selling at NVDA and $13.1M at CSCO — the people closest to the business were sellers. The completed 60-month backtest returned 14.9% over 191 trades with a 10.3% worst drawdown (the 12-month window lost 1.3%, so the edge is regime-dependent), but no robust parameter setup was formally recommended because the sensitivity evaluation exceeded its time budget. Today the strategy has literally nothing to buy: NVDA closed at $216.12 versus the required 9-day average of $222.30 and an RSI of 31.8 versus the required above-45, with $210 as the tactical invalidation level. The verdict that flips this idea is a strong NVDA fiscal Q3 print in late November that extends the $35B order thesis and repairs momentum — until then, wait.

Trade now: on the sidelines while NVDA rebuilds its trigger

**Nothing to buy today.** NVDA closed at $216.12, which is $6.18 below the 9-day moving average of $222.30 that the primary entry rule requires. The 14-day RSI sits at 31.8, well short of the required level above 45 — the biggest gap to close. The MACD line is already curling toward its signal line, so that condition is near, and price is holding above the nearest support at $210. In plain terms: two of the entry conditions are close, and the momentum gauge is the one that is not. The same setup applies to the CSCO leg: CSCO closed at $106.50 versus its 9-day average of $111.06, with an RSI of 30.8 against the same above-45 requirement. Note that both tickers currently sit below their 9-day averages — the exact condition that fires the exit rule — so even an existing position would be flagged for review, not held. "Wait" here means literally that: no new longs until price closes above the 9-day average *and* the RSI is back above 45 *and* the MACD line crosses above its signal line. The completed backtest supports waiting rather than forcing it. Over 60 months the pair strategy traded 191 times, won 47.6% of trades, returned 14.9% with a worst peak-to-trough drawdown of 10.3%; the 24-month window did better on return (15.7% across 86 trades). Position sizing is fixed-risk at roughly 2.5% per trade with a hard 2.5% stop-loss and a 5% take-profit, capped at 25% of the account per position. Once the trigger fires, the reward-to-risk on any new entry is about 2:1 by construction — but it only exists once the entry conditions line up. One caution on the tape: NVDA is 9.7% below its range high and 31% above its range low, with the nearest resistance at $216.83 — right where it closed. A bounce from here would still need to clear the $222 average before the strategy can act.

Orders, Not Hype: The Fundamentals Back the Fear-Gap Trade

The idea's core claim — that real orders are outpacing sentiment — is the easiest part of this thesis to test against the numbers, and Nvidia's latest reported quarter (fiscal Q4, period ending 2026-01-25) supports it directly. Revenue of $68.1B represents 73.2% growth year over year, and the full fiscal year came in at roughly $215.9B. Net income of $43.0B on a 63.1% net margin is not what a hype cycle produces when the crowd turns fearful — it is what a company with pricing power produces. Nvidia's operating margin of 65.0% sits in the 99.3rd percentile among 899 Information Technology peers, and free cash flow of $34.9B ranks in the 99.5th percentile. The thesis argues that demand, not narrative, is the driver; that is exactly what these percentile rankings show relative to the sector. The Reuters piece on the $35B order from ByteDance, Alibaba, and Tencent gives the bullish case a forward-looking catalyst on top of the backward-looking fundamentals. Per that report, three of China's largest tech firms placed over $35 billion worth of Nvidia chip orders — cash-backed demand from customers who could, in principle, buy elsewhere. In a value-oriented profile where the question is whether AI infrastructure spending pays off, an order book of that size is a concrete answer to the 'will anyone pay for this?' doubt the market has been pricing. The Cisco comparison, per the MarketWatch piece citing Tom Lee, adds a historical frame: past technology panics eventually resolved in favor of quality suppliers. Cisco's own numbers today partly vindicate that view of quality surviving the panic — its fiscal Q4 (ended 2026-07-25) revenue of $17.3B grew 17.6% year over year, free cash flow of $5.0B ranks in the 99.0th peer percentile, and it has raised its dividend steadily from $1.51 per share in 2022 toward $1.67 in 2026, roughly 2.5% annual growth. The quality-supplier thesis is not just a narrative; the companies it names are showing the cash generation to match. On the backtest evidence itself — this idea's realized, completed test of the NVDA/CSCO rule set returned 14.9% over 60 months across 191 trades, with a 47.6% win rate and a worst peak-to-trough drawdown of 10.3%. The 24-month window performed similarly (15.7%…

Scores

• Conviction score breakdown: 53
• Thesis support: 72
• Trade readiness: 30
• Risk quality: 55
• Backtest evidence: 62
• Fundamentals trend: 45

Watch items

• NVDA — RSI (14)
• NVDA — Close vs 9-day EMA
• NVDA — Nearest support level
• NVDA — Fiscal Q3 earnings report
• NVDA — Net insider open-market activity (next 13F/ownership filing)
• NVDA — Free cash flow, quarter over quarter
• CSCO — RSI (14)
• CSCO — Close vs 9-day EMA
• CSCO — Fiscal Q1 earnings report
• CSCO — Net insider open-market activity (next ownership filing)