An $18.6 billion warning knocked Nvidia, Palantir and Meta down together, which smells like indiscriminate selling of the whole AI trade rather than a problem at any one company. Push-back pieces argue the bear case on Nvidia is factually weak, and the CE

Verdict: A credible dip-buy thesis, but the entry rules have never fired — wait

The thesis — that the $18.6B warning knocked NVDA, META, MU, and PLTR down together as indiscriminate selling rather than company damage — has real fundamental support: MU just grew revenue 73.8% sequentially to $41.5B with an 84.6% gross margin, NVDA posted $96.2B in revenue up 17.9%, and PLTR compounded revenue 18.6% with an 84.7% gross margin and zero debt-to-equity. The strongest counterpoint is insider behavior: filings for the period ended June 30, 2026 show net open-market selling at all four names, including roughly $565.4M out of NVDA and $231.1M out of MU, and META's latest quarter showed free cash flow down 86.8% to $1.7B — signs the repricing may not be indiscriminate. The disciplined answer today is wait: the four-condition entry (band stretch, trend-line reclaim, RSI crossing back above 42, trend strength above 18) has produced zero triggers across 1,236 daily bars over five years, and even the closest name, PLTR, sits only near its $172.38 trend line and 42 RSI threshold while its trend strength of 0.37 is nowhere near the 18 minimum. A confirmed full entry on any of the four — especially against a market-wide scare — would flip this into an actionable buy; continued insider selling into a dip would argue the thesis is wrong.

Trade now: the rebound entry has not fired — here is exactly what would put it on

**This is a waiting setup, not an active signal.** The strategy's four-condition long entry — a stretch condition (lower Bollinger band below 50), a reclaim of the 20-day trend line, RSI crossing back above 42, and ADX above 18 — has not triggered on any of the four names. Only the ADX condition is live everywhere; the rest are not close. "Wait" here means: hold no position, check these levels daily, and act only when a full entry prints. **Where the four names stand.** NVDA closed at $232.65, trading above its 20-day trend proxy of $220.69 but with RSI at 67.6 versus the sub-42 stretch needed — so it is the opposite of the dip-buy state the strategy hunts. MU is the furthest away: at $1,017, some $66 above its trend line of $951, with RSI at 69.6. META at $617 is roughly $30 above its $587 trend line with RSI at 77.5. PLTR at $170.3 is the closest: it sits just $2.08 below its $172.38 trend line and its RSI of 43.6 is only 1.6 points above the 42 trigger — the first name to check each morning — but its ADX of 0.37 is nowhere near the 18 minimum, so it cannot trigger yet either. **Risk framing if an entry does fire.** The strategy carries a fixed per-position stop at a 2.4% loss and a take-profit at a 4.8% gain, an effective 2:1 reward-to-risk, with positions capped at 25% of the book and sized so a stop-out costs about 2.4% of equity. Those exits only matter after a valid entry, so do not front-run them today: buying now means buying without the strategy's invalidation line, in names where insiders were net open-market sellers in the June 2026 filings across the board (roughly $33.9M out at META, $231.1M at MU, $565.4M at NVDA, $136.6M at PLTR). **What would change today's answer.** A sharp one-day risk-off move — the very kind the thesis expects after an $18.6B warning shakes the AI complex — could compress all four toward their trend lines and stretch conditions quickly. That is when this turns actionable. Until one of the four entry rule sets fully fires, the disciplined move is to do nothing and let the pullback come to you.

The Bull Case: These Four Leaders Are Priced for a Scare, Not a Break

The thesis is that the $18.6 billion warning knocked all four names down together, which reads as indiscriminate selling rather than company-specific damage — and the fundamentals mostly support treating this as a scare, not a repricing. NVDA just posted $96.2B in quarterly revenue for the period ended July 26, 2026, up 17.9% sequentially, with a 75.0% gross margin and 66.2% operating margin, keeping it near the top of its peer set. On a full-year basis ended January 25, 2026, NVDA generated $34.9B in free cash flow with a 73.2% year-over-year revenue growth rate. The push-back pieces argue the bear case on NVDA is factually weak, and the CEO's comments on AGI point to demand that keeps outrunning supply. MU is the strongest fundamental turnaround in the group. For the quarter ended May 28, 2026, revenue surged 73.8% sequentially to $41.5B, net income more than doubled to $28.2B, gross margin expanded from 74.4% to 84.6%, operating margin jumped from 67.6% to 80.4%, and free cash flow spiked to $17.6B from $5.5B. The MarketWatch piece argues memory chips have come to rule the AI boom and that Micron's reign could be here to stay. Debt-to-equity fell 61.3% to 0.051 over the same period — a genuine balance sheet improvement alongside the revenue surge. PLTR is compounding steadily rather than explosively. For the quarter ended June 30, 2026, revenue grew 18.6% sequentially to $1.9B, net income rose 22.0% to $1.1B, and operating cash flow climbed 35.3% to $1.2B. Gross margin remains elite at 84.7%, and free cash flow of $1.2B puts it in the top few percent of its peer set. The balance sheet has zero debt-to-equity. META is the weakest of the four on the latest quarter but still elite on a full-year basis: FY2025 revenue of $201B grew 166% year-over-year with a 41.4% operating margin, $46.1B in free cash flow, and a 27.8% return on equity. The latest quarter's decline looks more like deliberate capex and spending shifts than a demand collapse, which is consistent with the thesis that this is a scare, not a structural break.

What the 'indiscriminate selling' argument leaves out

The bear case starts with the fact that this rebound thesis has never fired. The entry conditions were evaluated on 1,236 daily bars across 60, 24, and 12-month windows and produced zero entries — a rule demanding a specific pullback-plus-reclaim alignment did not occur once in five years. That does not invalidate the idea, but it means the market rarely hands out this setup: the entries are a watch-list, and nothing in the realized record…

Scores

• Conviction score breakdown: 47
• Thesis support: 70
• Trade readiness: 25
• Risk quality: 45
• Trigger proximity: 30
• Fundamentals trend: 65

Watch items

• PLTR — RSI (14)
• PLTR — Price vs EMA (20)
• PLTR — ADX (14)
• NVDA — RSI (14)
• META — Price vs EMA (20)
• MU — Price vs EMA (20)
• NVDA — Insider net open-market flow
• META — Free cash flow (next quarterly report)