Verdict: KEEP at 0.56% benchmark weight. Doorway state — two patterns compete with roughly equal probability, and at this weight the math says hold your nose and keep it. Removing the most oversold name in the basket at 13x GAAP earnings costs more in expected regret than it saves in expected drag.

But I want to be honest: this is 52% confidence. Almost a coin flip. The bear case is real, structural, and not fully resolved. I'm keeping it because the weight is small and the valuation floor is close, not because I have conviction.

The Filtration Question

Does ADBE drag on QQQ over the next 15 weeks? Maybe. The trailing 250-day idiosyncratic alpha is -57.4% annualized — that's the worst in any software name I've run through the regression. The decline is accelerating: -12% first quarter, -14% second quarter, -20% most recent quarter. That's not mean-reverting behavior. That's a stock in structural repricing.

But the stock sits at RSI 16.7 — the most oversold reading in the basket — trading at $241, just 3% above its 52-week low. GAAP forward P/E is ≈13.5x on 12% revenue growth. At some point, the price reflects the bad news. The question is whether that point is here or at $200.

I don't know. And at 0.56% weight, I don't need to know.

What Just Happened

Three things hit simultaneously on March 12:

1. Narayen announced he's leaving. After 18 years as CEO, Shantanu Narayen notified the Board on March 9 of his decision to transition. He stays until a successor is named, then moves to Board Chair. Frank Calderoni (Lead Independent Director) chairs the search committee. They're looking internal and external. Timeline: "a few months."

The timing is terrible. Adobe is navigating the most important technology transition since desktop-to-cloud, and the architect of that prior transition is walking away. Every CEO succession creates a 3-6 month uncertainty premium — but this one compounds with the AI disruption narrative. The new CEO will either be a "transformation" hire (aggressive AI pivot, potentially dilutive) or a "steward" (maintain trajectory, potentially uninspiring). Market doesn't know which, and won't until the announcement.

Historical analogs: Nadella at Microsoft and Cook at Apple were positive — but those were during strong business momentum. Apotheker at HP and the revolving door at Yahoo were negative — during strategic confusion. Adobe's situation is closer to the latter bucket than most shareholders want to admit.

2. Q1 beat, but Q2 guided down sequentially. Revenue hit $6.40B (+12% YoY), subscription revenue $6.20B (+13%), non-GAAP EPS $6.06 (+22% YoY). Record Q1 operating cash flow of $2.96B. The quarter was clean.

But Q2 guidance tells a different story: non-GAAP EPS guided $5.80-5.85, down 3.5-4.3% sequentially despite revenue guided slightly up. Non-GAAP operating margin compresses 300 basis points, from ≈47.5% to ≈44.5%. Part of this is Summit timing. Part is likely elevated retention grants — anti-dilutive shares (underwater options/RSUs) jumped from 2.6M to 6.3M YoY. When your equity comp is underwater, you need to issue more of it.

And then there's the tax rate. GAAP effective rate spiked from 17.0% to 22.0% — a 500 bps jump driven by "net tax expense related to stock-based compensation." This is a reflexive loop: lower stock price means less tax benefit from exercised SBC, which means higher taxes, which means lower EPS, which means lower stock price. The feedback loop runs until the stock stabilizes.

3. FY2026 targets reaffirmed. Despite the CEO change, Adobe held annual guidance. This prevents panic but also means Q2-Q4 must deliver to plan while the company searches for new leadership.

The Bear Case (Why This Stock Is Down 39%)

The market isn't stupid. Adobe's decline has reasons.

AI displacement is structural. The Q1 10-Q added explicit risk language about "generative and agentic AI solutions" — the first time Adobe named multi-modal and agentic AI as competitive vectors. This isn't boilerplate; it's a company acknowledging the threat in the one document where executives face personal liability for material misstatement.

Subscriber metrics are a black hole. Three full quarters of CC Pro global pricing and Adobe has disclosed exactly zero subscriber-level data: no paid subscriber counts (stopped reporting ≈2019), no conversion rates, no CC Pro acceptance vs. churn, no ARPU by plan, no generative credit economics. Revenue grew 12% — but we can't tell if that's organic volume growth or a pricing increase masking churn. Management's refusal to disclose isn't an oversight after three quarters. It's a choice. When management hides a metric, the metric is usually bad.

Segment consolidation killed transparency. Starting Q1 FY2026, Adobe reports as a single segment. No more Digital Media vs. Digital Experience breakout. No more segment operating margins. The investor can no longer separately track whether Creative Cloud standalone is decelerating faster than the blended number. The official reason — "unified selling motions and integrated product innovation" — is a corporate euphemism for "we don't want you to see the pieces."

The sell-side is turning. Goldman initiated SELL at $220. Argus downgraded from Buy to Hold. Four analyst downgrades in 8 weeks heading into Q1 earnings. Consensus is NEUTRAL (44% bullish, 10% bearish) — not the profile of a $98B company the street is excited about.

The Bull Case (Why I'm Not Removing)

Adobe Stock is only $450M. Narayen disclosed on the Q1 call that Adobe Stock is approximately $450M of book of business. Excluding it, ARR growth would have been 11.2% instead of 10.9%. That's a ≈30 basis point drag — manageable, not existential. The AI self-cannibalization thesis (Firefly replacing stock imagery) is real but the financial magnitude is at the low end of what bears feared.

AI-first revenue is growing. AI-first ARR tripled year-over-year. Firefly alone is at ≈$250M ARR. GenStudio and AEP apps grew 30%. Combined AI-first book of business is heading toward $1B. This doesn't offset the bear narrative on its own, but it shows the company isn't standing still.

Creative freemium MAU is 80M, up 50% YoY. The top of funnel is healthy. Users are coming. The conversion rate is unknown (that's the gap), but you need top-of-funnel growth before you can have conversion, and they have it.

The balance sheet is a fortress. $6.89B cash, $6.23B debt, $12B annualized OCF, $37M quarterly capex. This company generates essentially all its operating cash flow as free cash flow. They can buy back ≈8% of shares outstanding per year at current prices. There is zero going-concern risk, zero liquidity risk, zero covenant risk.

Valuation is at a floor. GAAP forward P/E is ≈13.5x. The non-GAAP figure of 9.14x that screams from every terminal is misleading — it strips out $2B+/year of stock-based compensation, which is a real economic cost to shareholders (dilution, or buyback drag to offset it). But even 13.5x on 12% revenue growth implies a PEG under 1.2. Mature SaaS with 5-8% growth typically trades 12-15x. The market is pricing ADBE as if growth is about to collapse to low single digits — and Q1 says it hasn't happened yet.

Factor Regression

ADBE = -57.4% alpha + 1.72*SPY + 0.25*XLK - 1.19*MTUM + epsilon
Idio variance: 67.8%  |  Idio vol: 25.5%  |  R-squared: 32.2%

The regression tells three stories:

First, the -57.4% trailing alpha. After removing market, tech, and momentum factors, ADBE has been declining at 57% annualized for company-specific reasons. This is severe. It's not factor rotation — it's the market repricing Adobe specifically.

Second, the 67.8% idiosyncratic variance is below the 75% target. ADBE is partially factor-driven — high beta (1.72) means market moves explain a lot of the variance. For the QQQ hedge this partly cancels, but it means less of the stock's movement is addressable through company-specific research.

Third, the negative momentum loading (-1.19). ADBE moves opposite to the momentum factor — classic loser profile. This means if momentum reverses (market rotation from winners to losers), ADBE would benefit disproportionately. That's a potential 15-week tailwind if rotation occurs.

The 15-Week Math

Three catalysts sit inside the window:

CEO successor (weeks 1-15). "A few months" means 50% probability of announcement within our horizon. Binary: good hire catalyzes re-rating, prolonged search or weak hire extends the decline.

Semrush acquisition close (weeks 4-8). $1.9B cash, already guided. Low-impact unless integration reveals problems.

Q2 FY2026 earnings, June 11 (week ≈11). The big one. C&MP subscription revenue guided $4.41-4.44B. If YoY growth holds at 10%+, relief rally from these levels. If it breaks below 10%, the slow erosion thesis graduates from pattern to fact.

Scenario distribution:

Expected value: $247, or +2.6% from current. At 0.56% weight: +1.5 bps contribution.

Edge Audit

I have no informational edge on this name. $98B market cap, 39 analysts, covered by Goldman, JP Morgan, Morgan Stanley. Everything I know — the subscriber opacity, the CEO transition, the AI displacement — is visible to every institutional investor in the market.

The only "edge" is recognizing that extreme oversold conditions at valuation floors tend to mean-revert over 3-4 month horizons. That's a statistical pattern, not an insight. I'm betting on base rates, not research.

The counterparty at $241 is likely momentum funds and tax-loss sellers — uninformed relative to fundamental value but informed relative to flow dynamics. Against Goldman's SELL desk at $220 target, I have no reason to believe my estimate is better.

What Would Flip to REMOVE

• C&MP growth below 8% (not 10% — the market expects some deceleration)
• CEO search extends past 6 months with no frontrunner
• Key product exec departures (Wadhwani, Chakravarthy)
• Short interest spikes above 5% (currently 3.6%)
• Another segment transparency reduction or metric withdrawal

Bottom Line

ADBE is the kind of name that makes filtration hard. The bear evidence stacks up — cumulative Bayes says bearish, the alpha is deeply negative, the CEO is leaving, the metrics are hidden. But the stock has already moved -39%, the valuation is compressed to levels that imply near-zero growth, and the position is too small to matter.

If I had to bet real money on direction over 15 weeks, I'd give slight odds to a bounce — not because I believe the thesis is wrong, but because the price has already incorporated most of what the thesis says. At 0.56%, the expected cost of being wrong in either direction is single-digit basis points.

Keep it. Watch the earnings. If C&MP breaks below 10%, revisit.