The AI trade is splitting into two camps: companies actually making money from AI today versus stories priced on future promises. Fitch's warning that an AI bust could sink stocks 35% means the speculative end of the trade is fragile, especially with a ra

Verdict: The profitable-AI names deserve your watchlist, not your capital — yet

**Verdict: a strong thesis waiting for a real entry — stay on watch, not in the market.** The strongest point for the trade is the quality of the underlying names: Nvidia's most recent fiscal year showed $215.9B of revenue up 65.5% with $96.7B of free cash flow (top 1% of the IT sector), and Palantir's operating margin climbed to 47.1% in the quarter ended June 30, 2026, up from 46.2% — this is AI demand being monetized, not promised. The strongest point against is that the setup has never actually triggered: across 185 daily bars in the most recent nine-month window there were zero entries, and no robust parameter setup was established — plus the ownership filings for the period ended June 30, 2026 (a delayed disclosure, not a current read) show net open-market insider selling of roughly $565.4M at Nvidia, $136.6M at Palantir, and $110K at Adobe. Adobe is the soft spot in the profitable camp: net margin fell 3.7 points to 25.9% and operating margin dropped 4.0 points to 33.8% quarter over quarter, with an interim CFO in place during the AI pivot. Right now Nvidia is closest to a full entry state — price above its 50-day average and RSI at 45.4 versus the 45 threshold — but trend strength at 10.0 sits well below the required 20. The single fact that would flip this verdict to actionable: a daily close where any of the three satisfies all three entry conditions (price above the 50-day average, RSI above 45, and trend strength above 20) together with the next quarterly prints confirming Adobe's margin dip is one-off rather than trend.

Trade now: no entries yet — three trigger sets, none fully armed

Nothing to buy today. The strategy waits for a pullback-and-reclaim state on each name — price back above the 50-day average, RSI (14) crossing above 45, and trend strength (ADX, 14) above 20 — and right now no ticker satisfies all three at once. Nvidia is closest: price is $219.29 versus a 50-day average of $215.38 (already above), RSI (14) sits at 45.4 versus the 45 threshold, but trend strength at 10.0 is well under the 20 floor. Palantir at $167.23 is above its 50-day average of $159.76 and trend strength at 25.3 clears the bar, but momentum at 41.1 still needs to reach 45. Adobe is furthest away: at $252.23 the stock sits below its 50-day average of $257.32, momentum at 32.7 is 12.3 points from the trigger, and it is 2.6% under the first resistance shelf at $256.43. Risk is defined mechanically rather than by conviction: the stop sits at a 2.5% adverse move and the target at a 4.9% gain, roughly 2:1 reward to risk, with positions capped at 25% of the book. The exit side is far from live — every RSI (14) reading is at least 29 points below the overbought exit at 75, so any future position would be governed by the price stop and target first. One factual scope note: the frozen rule set could not be evaluated historically because daily data coverage for Palantir remained incomplete, and no robust parameter setup was established as a result; the author requested bounded optimization with the thesis and holdout preserved. Treat this strictly as a watch-list setup — "wait" means setting alerts at the levels in the table below and doing nothing until a full entry state prints on a daily close.

The Profitable-AI Camp Has the Receipts

The thesis splits the AI trade into companies earning from AI today versus stories priced on promises. The numbers back that split. NVIDIA's most recent fiscal year (ended January 2026) shows $215.9B of revenue, up 65.5% year over year, with $120.1B of net income and $96.7B of free cash flow — free cash flow that ranks in the top 1% of roughly 790 Information Technology peers. That is demand being monetized now, not forecast. Palantir is the second leg of the argument. Its FY2025 revenue of $4.48B grew 56.2%, gross margin sits at 82.4%, and the quarterly series shows steady margin expansion into 2026 — operating margin reached 47.1% in the quarter ended June 30, 2026, up from 46.2% the prior quarter, while free cash flow hit $1.2B that same quarter. The idea's demand-confirmation claim about the Nvidia/Palantir buildout (per the September 10 Yahoo Finance piece on speeding up the AI buildout) is consistent with those filings. Adobe is the anchor of the profitable-camp case. The September 10, 2026 report of a record quarter and raised guidance is corroborated by the reported data: FY2025 revenue of $23.8B, up 10.5% year over year, 89.3% gross margin (97th percentile among peers), and $9.85B of free cash flow in the top 1% of the sector. Adobe also repurchased shares down to 399M shares outstanding in its latest quarter, from 406M a quarter earlier — cash generation funding shareholder returns rather than dilution. On the entry mechanics: the strategy's rules were evaluated on real daily bars but did not open an entry — across 185 evaluated bars in the most recent nine-month window there were no triggers, and earlier windows could not complete evaluation. This is a watch-list setup, not an active signal, which arguably suits the thesis: the idea is about quality you want to own on pullbacks, and the setup explicitly waits for a pullback toward a medium-term trend line with momentum turning back up. The research author requested bounded optimization after zero entries rather than shipping an untested trigger, with no robust setup yet…

Scores

• Conviction score breakdown: 45
• Thesis support: 65
• Trade readiness: 25
• Risk quality: 50
• Trigger proximity: 30
• Fundamentals trend: 55

Watch items

• NVDA — RSI (14)
• NVDA — ADX (14)
• PLTR — RSI (14)
• ADBE — RSI (14)
• ADBE — Price vs 50-day EMA
• ADBE — Nearest support shelf
• NVDA — Insider net open-market activity
• ADBE — Net margin (Q/Q)