OpenAI trims its raise and Anthropic's IPO hits turbulence — short the richest AI growth names

Verdict: the AI-funding-crack short has a narrative hook but its own numbers argue against it — wait for the rally-tag

The verdict on this bearish AI-narrative idea, over a next-quarter horizon, is that the catalyst story has real hooks but the trade is not yet justified. The strongest point for the thesis is filed: both shorted names show net open-market insider selling in the provisional ownership records for the period ended September 30, 2026 — roughly $146.3M at Palantir and $8.0M at Super Micro — though those totals are provisional, not final. The strongest point against is the fundamentals: Palantir's latest quarter (ended June 30, 2026) posted $1.94B of revenue, up 18.6% quarter over quarter, with a net margin near 55% and $1.20B of free cash flow, and Super Micro swung back to positive $719M of free cash flow with gross margin recovering to about 17.5% from about 10% — both self-funding, which undercuts the idea's core premise that these are promises-over-profits names. The completed backtest also shows a thin edge: 0.6% total return on 33 trades over 60 months at a 45.5% win rate with a 22.4% maximum drawdown, and exits were filled on daily bars, so the record is coarse. The entry itself is only partially triggered — Palantir's high has not yet tagged its 50-day average near $187 (close $184.11) — and its RSI of 30.4 sits just 0.4 points above the oversold exit, so chasing here risks an immediate stop or profit-take. We wait; a confirmed rally-tag entry or a renewed negative-cash-flow quarter at Super Micro would change the assessment.

Trade now

**Wait for the bounce — do not chase the short here.** The idea argues that AI-linked names trading on promises rather than profits are the most exposed as private-market AI funding tightens, and it expresses that by shorting the two richest AI-adjacent growth names (PLTR and SMCI) rather than the index. That business thesis may or may not be right — notably, PLTR's latest reported quarter showed revenue up 18.6% quarter over quarter to $1.94B with a net margin near 55%, which is not a company with no current profits — but the entry question is separate, and today's entry conditions are only partially met. The strategy's entry needs a specific rally-into-resistance pattern: the 50-day average below the 200-day average (true for both names), price closing below the 200-day average, the day's high reaching the 50-day average while the close finishes below the 20-day average, and RSI (a 0–100 momentum reading) below 50. As of the October 31 observation (daily bars), PLTR's high has not yet tagged its 50-day average near $187 — price closed at $184.11, about 1.5% short of that trigger — so the setup is armed but not fired. The 45-day maximum holding period, the oversold-RSI exit near 30, and the hard stop at a 3% loss against a 10% profit cap define the trade's shape once it triggers. The downside that matters most: shorting PLTR means fighting a stock whose worst two-year drawdown was 48.2% but whose annualized return over the same window was roughly 75%, and the backtest itself shows a 22.4% maximum drawdown and a 45.5% win rate over 33 trades in 60 months — a grind, not a guarantee. ### Technical details Entry rule set (per name, daily bars): short when (1) the 50-day average is below the 200-day average; (2) the close is below the 200-day average; (3) the session high reaches the 50-day average — the rally-tag condition; (4) the close is below the 20-day average; (5) 14-day RSI is below 50. Current readings, observed October 31: PLTR — 50-day average $186.90 vs. 200-day average $153.58 (condition met), close $184.11 vs. 200-day average $153.58 (met), high vs. 50-day average not yet met (price is $2.79 below, roughly 1.5%), close below the 20-day average $197.25 (met), RSI 30.4 (met). SMCI — 50-day average $38.17 vs. 200-day average $32.57 (met), close $32.13 below the 200-day average (met, by $0.44), price $6.04 below the 50-day tag level (roughly 16% short of triggering), close below the 20-day average $36.76 (met), RSI 34.7 (met). Exits: time exit after 45 trading days; profit-take on a close at or below the nearest support level (PLTR $184.45, SMCI $32.00) or RSI below 30 (PLTR is only 0.4 points above this — an entry near current levels would risk an immediate oversold exit); stop-loss on a close above the 50-day average (PLTR $186.90, roughly 1.5% above spot; SMCI $38.17, roughly 19% above spot) or a hard −3% position stop; profit cap at +10%. Position sizing is fixed-risk at 2% per trade, max 25% per position. Reward-to-risk is therefore roughly 10:3 by the hard brackets, but the 50-day stop on PLTR sits just above entry and RSI at 30.4 sits just above its profit-taking threshold — which is exactly why the ruleset wants a bounce toward the 50-day average first. QQQ appears in the watch list only as regime context; its price of $685.42 sits just above its 200-day average of $679.39, so the index has not confirmed the same weak regime. One data caveat: the strategy executes against the two single names, and exits were filled on daily bars rather than intraday, so backtest fill quality is approximate.

The funding-squeeze case for shorting the richest AI names

The bearish idea is supported where it is most literal: the two shorted names show insider selling, and Super Micro shows a cash engine that just had a catastrophic year. Per CNBC's October 31 report on OpenAI trimming its funding target by $10 billion and the piece on Anthropic's rougher IPO path, the thesis argues AI financing confidence is cracking; the filings and fundamentals supplied give that narrative some hard-numbers hooks. ### Insiders are net sellers into the squeeze Ownership filings for the period ended September 30, 2026 show net open-market insider selling of roughly $146.3 million at Palantir (569 holders reporting) and $8 million at Super Micro. Both records are marked provisional, so the totals may still move, but the direction is unambiguous: the people closest to both businesses were reducing exposure while the AI-narrative-repricing thesis was being formed. For a short idea premised on insiders and holders blinking first, that is the single most on-thesis piece of filed evidence. ### Super Micro's cash engine broke in FY2026 Super Micro's fiscal 2026 (ended June 30) free cash flow was negative $6.97 billion, driven by a single quarter ending March 2026 at negative $6.70 billion, and its debt-to-equity ratio jumped to 0.54 in that quarter from 0.03 two quarters earlier, settling at 0.28 by fiscal year-end. A hardware business funding $39.1 billion of annual revenue (up 93% year over year) with balance-sheet debt rather than customer cash is exactly the profile the idea says is exposed when financing tightens — working-capital strain is…

Scores

• Conviction score breakdown: 37
• Thesis support: 50
• Trade readiness: 35
• Risk quality: 40
• Backtest evidence: 30
• Fundamentals trend: 30

Watch items

• PLTR — Session high vs SMA(50) entry tag
• PLTR — RSI (14)
• SMCI — Session high vs SMA(50) entry tag
• QQQ — Close vs SMA(200)
• PLTR — Close vs SMA(50) stop level