Adobe did everything right — beat earnings and raised guidance — and the stock still fell, because investors now punish any AI story without clear payback. Barron's is pointing to multiple late-stage bubble warning signs in AI stocks, and Michael Burry is

Verdict: wait for the ADBE breakdown to confirm before shorting the AI software basket

The idea's core read — that the market now punishes AI software names without clear payback, per the Yahoo Finance piece on Adobe's beat-and-raise being sold — is well supported by the filed data: ServiceNow's operating margin collapsed from 13.3% to about 4.0% and Adobe's fell from roughly 37.8% to about 33.8% in their latest quarters, and the June 30, 2026 ownership filings show net open-market insider selling at all four names, led by roughly $136.6M net sold at Palantir. But the bear case is weakest exactly where it matters most for the basket: PLTR grew revenue 56.2% in fiscal 2025 with a 36.3% net margin and rising operating margin, so shorting the group's fastest grower with its most profitable name included is the strongest point against. On the live setup, only ADBE is close — price is $5.09 below the 50-day EMA and needs just a 0.5% close below $250.88 to trigger — while CRM is fully in exit territory with price $37.83 above its EMA and RSI of 58.9. The backtest is honest but choppy: +134% over 60 months on 93 trades, yet -10.3% over the trailing 12 months with a 37.1% drawdown and a historical worst of 47.8%, so sizing for the drawdown regime matters more than the headline. The verdict is to wait for the ADBE support break to confirm, and to treat any print showing margin re-expansion at ADBE or NOW as a thesis invalidation candidate. If ADBE reclaims $257.32 with RSI back above 55, the signal flips to exit and the short-side framing is wrong.

Trade now: ADBE is one support break from triggering — the rest of the basket is still waiting

The live setup is closest on **ADBE**. At $252.23, all three filter conditions are already met: price is $5.09 below the 50-day EMA at $257.32, RSI (14) is 32.7 (below the 50 threshold), and ADX (14) is 36.7 (above 20). The only condition outstanding on the crossover entry is a daily close **below the first support level at $250.88** — just 0.5% below the current price. A second, signal-only entry on ADBE (price below EMA, RSI below 50, ADX above 20, no support cross required) is **already fully met**, so that setup is live today. Risk is mechanical: each position carries a **stop at 2.5% below entry** and a **take-profit at 5.0% above entry**, an effective reward-to-risk of roughly **2:1**, with position risk capped at 2.5% of equity per trade and a 25% maximum position size. If the support-cross entry fires near $250.88, that puts the stop around $244.49 and the target around $263.42; the EMA reclaim near $257.32 is the first signal-based exit threshold, alongside RSI (14) back above 55. "Wait" means something specific for the other three names. **CRM** at $247.72 is nowhere near entry — it needs a close below its 50-day EMA at $209.89 (about 15% lower) and RSI at or below 50 (now 58.9); in fact CRM is currently in exit territory, with both the price and RSI exit conditions met. **NOW** at $132.53 has RSI below 50 (46.8) but sits $9.50 above its EMA at $123.03 and its ADX of 11.4 is well under 20, so no entry there yet. **PLTR** at $167.23 is the next-closest watch: RSI (41.1) and ADX (25.3) already qualify, and price needs to fall $7.47 to its EMA at $159.76 to come in range. On evidence: the strategy is backtested over 60 months on this basket, producing 93 trades with a 43.0% win rate, a 134% total return, and a 47.8% maximum drawdown — but the most recent 12 months ran a -10.3% return with a 37.1% drawdown, so size for the drawdown regime, not just the headline. No robust nearby parameter setup was established by the sensitivity review, so trade the published rules as written.

The Bear Case Has Real Numbers Behind It

The idea's core claim — that Adobe could beat and raise and still fall because the market now demands AI payback — is exactly what the numbers support. On a fiscal year ended November 2025, Adobe grew revenue 10.5% year over year, which lands at only the 50th percentile among 788 Information Technology peers. That is the profile of a company whose growth is no longer justifying an AI narrative premium, and per the Yahoo Finance piece on Adobe and Salesforce, the market punished a beat-and-raise anyway. The most recent quarter makes it worse. In the quarter ended May 29, 2026, Adobe's net margin fell from 29.5% to 25.9% and operating margin from 37.8% to 33.8% quarter over quarter. ServiceNow is even more exposed on this axis: its Q2 (ended June 30, 2026) gross margin dropped 4.4 points to 70.7%, and operating margin collapsed from 13.3% to 4.1% — arguably the first hard evidence of AI-related cost pressure hitting the software P&L rather than just the narrative. Ownership posture reinforces the bear case. As of the 2026-06-30 filings, all four names show net open-market insider selling, led by Palantir at roughly $136.6 million of net sales and ServiceNow at $3.8 million. When insiders sell a crowded theme in size at the same time, the marginal buyer is gone — and per the Yahoo Finance report on Michael Burry, a well-known investor is publicly short Palantir while warning of a sharp repricing, and Barron's (September 11) flagged multiple late-stage bubble warning signs. The realized evidence is consistent with the thesis in the recent regime. The completed backtest on Adobe's daily bars returned -10.3% over the trailing 12 months across 22 trades with a 36.4% win rate, and the 24-month window returned -20.1% across 41 trades at a 36.6% win rate — the strategy made its money over the full 60 months (+134.1%, 43.0% win rate on 93 trades, with a worst peak-to-trough drawdown of 47.8%) but the recent-year regime is precisely the 'good news stops lifting the theme' environment the idea describes. Note the parameter-sensitivity review ended with no robust nearby setup established, so the published rules are the thesis-consistent ones.

The Group's Best Business Doesn't Look Like a Bubble

The strongest counterargument is…

Scores

• Conviction score breakdown: 56
• Thesis support: 75
• Trade readiness: 55
• Risk quality: 45
• Backtest evidence: 55
• Fundamentals trend: 50

Watch items

• ADBE — Daily close vs first support
• PLTR — Close vs 50-day EMA
• NOW — ADX (14)
• NOW — Close vs 50-day EMA
• CRM — Close vs 50-day EMA
• CRM — RSI (14)
• ADBE — Insider open-market net value
• PLTR — Insider open-market net value
• NOW — Operating margin, next quarterly report
• ADBE — Permanent CFO appointment