Anthropic's $3.5B IPO lands amid AI gloom while Korea funds its own AI — buy the dip in AI infrastructure

Verdict: the AI demand story is real — but wait for the dip-buying trigger to confirm

Over a near-term, event-driven horizon, this idea has a strong demand case but an unconfirmed entry: Anthropic's $3.5B Japanese IPO and South Korea's record AI budget land amid a sentiment trough while all three names still show accelerating fundamentals (NVDA revenue up 17.9% to $96.2B, MSFT operating margin near 49.8%, AMD net margin up 6.4 points to 19.9%). The strongest support is that the October selloff does not look fundamentals-driven — the most recent reported quarters show growth and margin expansion across the board. The strongest risk is that insiders at all three companies filed net open-market selling for the period ended September 30, 2026 (roughly $569M at NVDA, $186M at AMD, $52M at MSFT, provisional), and NVDA's free cash flow fell 56% quarter over quarter — a warning that the dip may be a rational repricing. The entry rules are not live today: only MSFT has RSI above 40, and it still closes about 1% below the required $530.66 level, while AMD and NVDA sit below the 40 RSI threshold. If MSFT closes back above $530.66 with the momentum cross intact, or Anthropic confirms the IPO date, the verdict changes; if the IPO slips or NVDA's next print shows another cash-flow collapse, the thesis weakens. Scope note: this exact entry setup was not backtested because MSFT and NVDA raw and adjusted price series use different bases, so no robust setup was established — this rests on live conditions and filed fundamentals.

Trade now

The business thesis is attractive but the entry conditions are not yet met, so the answer is: wait for the setup's own trigger, do not buy today. The idea argues that real AI demand — a $3.5B Anthropic IPO for Japanese investors and South Korea's record AI chip and data-center budget — is arriving into a sentiment trough after a brutal October for tech (per the idea's thesis). The reported numbers do not contradict that: NVDA's latest quarter (ended July 2026) showed revenue up 17.9% quarter over quarter to $96.2B with a roughly 75% gross margin, and MSFT's September quarter showed revenue up 2.5% to $92.3B with operating margin expanding to about 49.8%. AMD's June quarter revenue rose 12.5% to $11.5B with net margin jumping from 13.5% to 19.9%. The demand story has fundamental support. The trade timing, however, fails its own tests. On the daily charts (as of the November 2, 2026 observation), all three names sit below the level at which the momentum rule fires: AMD's RSI is 36.4 versus a 40 threshold, NVDA's is 37.1, and only MSFT at 46.4 has crossed — but MSFT also fails the price condition, closing at $525.35 versus the required level of about $530.66. AMD's entry is roughly $57 away from its price condition, the furthest of the three. ADX, which measures trend strength, is comfortably above its 15 threshold for all three, so the missing piece is purely the momentum cross plus price reclaiming its lower band. The horizon here is near-term and event-driven. "Waiting" means tracking the daily close: an entry becomes evaluable when the 14-day RSI crosses above 40, the 14-day ADX stays above 15, and the close moves back above the 20-day lower Bollinger band for a given ticker. The downside that matters is twofold: AMD's own trailing drawdown over the two-year lookback reached 47.8%, and filed ownership data (provisional, September 30, 2026 period) shows net insider open-market selling across all three names — roughly $569M at NVDA, $186M at AMD, and $52M at MSFT — a caution against assuming the dip is fully bought. One scope note: historical performance of this exact rule set was not measured because MSFT and NVDA raw and adjusted price series use different bases, so this assessment rests on live conditions and fundamentals, not backtested statistics.

Fundamentals Say the AI Money Is Real — Even as Sentiment Rolls Over

Takeaway: The idea's core claim — that real AI demand is accelerating while investor expectations are depressed — is supported by the hardest evidence available: all three names delivered accelerating revenue, expanding operating margins, and roughly $4.7B to $12.6B of fresh free cash flow in their most recent quarters, even as a brutal October and a major bank abandoning its big-tech trade (per the MarketWatch coverage) lowered the bar. ### NVDA's latest quarter is the strongest single proof of the demand thesis NVDA's fiscal Q2 (ended July 26, 2026) showed revenue of $96.2B, up 17.9% quarter over quarter from $81.6B, with net income of $59.7B and a 66.2% operating margin. That is the actual spending of AI infrastructure customers showing up on the top line — not a forecast. The trailing four quarters tell the same story: full-year revenue of $216B on the January 2026 fiscal year, with the latest annual snapshot showing 73.2% year-over-year revenue growth. If hyperscaler and sovereign spending were decelerating, these numbers would be bending down; instead the quarterly sequence ($81.6B → $96.2B) is still climbing. ### MSFT confirms the demand from the buyer side, not just the supplier side MSFT's fiscal Q1 (ended September 30, 2026) grew revenue to $92.3B (+2.5% quarter over quarter on top of $90.0B), with operating margin expanding 4.65 points to 49.8% and net income of $36.6B. Critically, MSFT's capex hit $35.8B in its fiscal Q4 (ended June 2026) — the 15th percentile of R&D intensity among 1,138 Information Technology peers (R&D at roughly 11.1% of revenue) alongside that capex means MSFT is one of the companies actually building the data centers the idea's sovereign-budget catalysts fund. When the largest buyer of AI chips keeps growing revenue and margin simultaneously, that is direct corroboration that end-demand is not a narrative. ### AMD shows the second-derivative is positive, exactly what a sentiment-trough entry needs AMD's fiscal Q2 2026 (ended June 27) grew revenue 12.5% quarter over quarter to $11.5B, lifted net margin 6.42 points to 19.9%, and raised operating margin 2.85 points to 17.3% — all while its debt-to-equity ratio actually improved to 0.048 from 0.050. AMD is also the 81.7th percentile of revenue growth among 1,131 IT peers. This matters because the thesis's 'buy the dip' mechanism requires the laggard name to have improving fundamentals, not deteriorating ones: AMD's gross margin has climbed from 50.2% (Q1 2025) to 53.8% (Q2 2026), meaning any re-rating toward MSFT/NVDA levels is being earned operationally, not just hoped for. ### The sentiment setup matches the fundamentals, not contradicts them The idea's catalyst stack — Anthropic's $3.5B Japanese IPO (per the Reuters report dated November 2, 2026) and South Korea's record sovereign AI budget (per CNBC, November 2, 2026) — lands while Citi is abandoning a big-tech trade (per MarketWatch, November 2, 2026). Buying quality names 'on the dip' is only attractive if the dip is not being driven by fundamentals. Here it isn't: the most recent reported quarters for all three names show revenue growth, margin expansion, and — for AMD — net income up 66.1% quarter over quarter. The thesis's premise that expectations have been reset while demand has not is directly consistent with the filed numbers. ### Scope note on the entry mechanics This idea's historical performance was not measured: the supplied backtest could not be computed because MSFT and NVDA raw and adjusted price series use different bases, and no robust parameter setup was established. What follows is therefore a…

Scores

• Conviction score breakdown: 61
• Thesis support: 78
• Trade readiness: 35
• Risk quality: 50
• Fundamentals trend: 82

Watch items

• MSFT — Close vs lower Bollinger (20)
• AMD — RSI (14)
• NVDA — RSI (14)
• NVDA — Insider net open-market activity (provisional, Sept 30 2026 period)