Verdict: KEEP | Weight: 0.70% | Confidence: 75%

The Game

INTU is rank #26 in QQQ at 0.70%. Removing it redistributes ≈1.4 bps per surviving name. The question: will Intuit underperform QQQ over 15 weeks badly enough to justify removing it from the basket?

The trailing alpha screams remove. The fundamentals scream keep. One of them is lying.

The Price Chart vs The Income Statement

INTU is down 47% from its $813 peak. RSI 31. Sitting at the 18th percentile of its 52-week range. Trailing 250-day regression: alpha -43.6% annualized, 77.5% idiosyncratic variance. Momentum loading -0.67. By the numbers, this is a momentum loser with deeply negative company-specific alpha.

Now read the income statement.

Q2 FY2026 (ended January 31, 2026): revenue $4,651M (+17.4% YoY). GAAP operating income $855M (+44.2%). GAAP EPS $2.49 (+48.2%). Non-GAAP EPS $4.15 (+25.0%). H1 FY2026: revenue $8,536M (+17.8%), GAAP EPS $4.09 (+71.8%).

Every major revenue line is accelerating or stable:

QBO Accounting has accelerated for six consecutive quarters: +17% → +21% → +22% → +23% → +25% → +24%. Mid-market contracts grew 50% quarter-over-quarter. QuickBooks Live (human expert + AI) grew 50% YoY.

The only line declining: Mailchimp. Revenue down slightly. CFO Aujla on the call: "All options are on the table, and we'll make sure we keep you all apprised as we narrow in." Read: divestiture. Mailchimp was a $12B acquisition with $4.2B of goodwill sitting on the balance sheet. It's the drag everyone can see. Cutting it loose would be a catalyst.

This divergence — deeply negative trailing alpha on accelerating fundamentals — is the SaaS AI narrative selloff in one chart. The market punished the entire software stack on the thesis that AI agents would eat subscription software. Intuit's numbers say the opposite: AI is accelerating revenue growth, expanding margins, and driving consumption.

The price chart is backward-looking. The 10-Q is forward-looking. For a 15-week filtration window, the 10-Q wins.

The March 16 Signal

On March 16, INTU filed an 8-K disclosing that its founder and the entire executive leadership team terminated all outstanding pre-scheduled 10b5-1 stock sales plans. Simultaneously, the company reiterated its intent to "substantially accelerate" share repurchases under a $3.5 billion remaining authorization.

Context: H1 FY2026 repurchases were $1.8B, up 40% YoY. At $432/share, the remaining $3.5B buys back approximately 8.1 million shares — 2.9% of the outstanding ≈278 million.

This is a Tier 1 signal. 10b5-1 plans are filed with the SEC. Cancelling them means executives chose to stop selling their own shares at $350-$432. Every single one of them. Simultaneously. While the company accelerates buybacks at what management clearly believes are depressed prices.

Executives go to jail for lying in SEC filings. They don't cancel selling plans as a marketing exercise.

Tax Season: The 15-Week Catalyst

Q3 FY2026 earnings report May 21 — dead center in our filtration window. Q3 (Feb-Apr) is peak TurboTax season.

The setup is unusually favorable:

IRS Direct File is shut down. The government's free tax filing alternative was suspended for the 2026 season after achieving only ≈1% adoption among 30 million eligible filers. One competitive threat eliminated.

No AI filing competitor exists. ChatGPT cannot e-file taxes. It has no IRS authorization, no state filing integrations, no audit support. OpenAI's partnership with H&R Block targets tax professionals, not DIY consumers. The prediction market on this — our own, 92% probability — that no frontier AI product launches consumer tax filing with IRS e-file authorization for the 2026 season. TurboTax's moat is regulatory, not technological.

Early season metrics are strong. CEO Goodarzi on the Q2 call: "We've seen enough of the tax season to know how it's going to play out." Through February 6, TurboTax revenue grew 12% while IRS returns were down over 5 points. Compare to last year: TurboTax +4% on IRS returns -8%. Massive relative improvement. Tax landing page and service center visitors hit 5.1 million through February 6 — exceeding the 4.2 million for the entire prior season.

Guidance for Q3: revenue growth +10%, non-GAAP EPS $12.45-$12.51. Given the beat pattern — four consecutive quarterly beats of +3% to +29% on EPS — and the strong early tax season data, this guide looks conservative.

What's Actually Growing

Mid-market disruption is real. QBO Advanced and Intuit Enterprise Suite revenue grew ≈40% in Q2. New IES contracts grew 50% QoQ. Sales team expanding 30% on "attractive LTV to CAC economics." Accounting firm partnerships accelerating — 1/3 of new contracts influenced by accountant recommendations in Q2, up 10 points from Q1. The construction-specific ERP vertical launched in February. This is a $90B TAM with 6% penetration.

AI agents are driving consumption, not replacing it. 3 million+ customers using agents (up from 2.8M four months prior), 85%+ repeat engagement. In January alone, AI accounting agents categorized 237 million transactions — more than half of all transactions that month. Business tax agent uncovering average $1,000+ in incremental deductions per customer.

The critical finding from management's testing: when AI and human intelligence are offered as a combined "done-for-you" experience, customers pay more, not less. QB Live grew 50% YoY. Customers with expert engagement show 22 points higher ecosystem attach. AI is an upsell vector, not a margin compressor.

The Anthropic and OpenAI partnerships are defensive moats, not risks. Intuit announced a multiyear partnership with Anthropic: Claude's Agent Builder will use Intuit's domain-specific models, Intuit capabilities come to Claude and Cowork users. All four apps launched in OpenAI's App Directory. CEO's framing is correct: "Data doesn't leave our four walls... we own the experience and the relationship, and we don't share in the economics." The LLM providers are distribution channels for Intuit, not competitors. Tax compliance and bookkeeping accuracy require regulatory authorization, proprietary data models, and liability frameworks that frontier labs neither have nor want.

The Red Flag: Credit Losses

QBO Capital is growing the loan book aggressively. H1 FY2026 originations: $2.6B (+73% YoY). But charge-offs: $75M (+108% YoY). Net portfolio: $1.6B. Allowance: $120M (+43% from Jan 2025 to Jan 2026).

The math: annualized net charge-off rate is approximately 8-9% on a $1.6B unsecured SMB loan portfolio. That's high. And charge-offs are growing at roughly double the rate of the loan book. Provision for credit losses in H1: $87M, representing 1.0% of total revenue.

Is this a thesis-breaker? Not today. It's a $87M drag on an $8.5B revenue base. But it's the number to watch. If macro deteriorates — and CFO Aujla flagged that micro businesses are seeing cash balances decline while mid-market and SMB remain stable — the charge-off rate could cross 12%, at which point it becomes material to margins.

The lending business is also partially de-risked: INTU sold $595M of originated loans in H1 FY2026 (up from $246M). They're originating and flipping, not just holding.

The Bear Case, Steelmanned

1. Negative momentum persists. MTUM loading -0.67. Momentum strategies are short INTU. If factor flow continues, mechanical selling pressure continues regardless of fundamentals. Momentum crashes happen, but so do momentum continuations.

2. Revenue growth is ARPC, not customers. QBO paying customers grew only 5% in FY2025. Q1 FY2026 customer growth "came below expectations." Revenue growth decomposes to roughly 55% pricing, 20-25% mix-shift, 20-25% customer additions. INTU discloses zero retention, NRR, or churn metrics in any SEC filing. Xero offers a free QBO migration tool. There's a ceiling to pricing power if the customer base isn't expanding.

3. 34 analysts cover this stock. No informational edge exists on a $120B market cap company with universal sell-side coverage. Everything we've found — the March 16 8-K, the credit loss trajectory, the tax season metrics — is visible to every institutional participant. Our counterparty is the entire informed market.

4. Valuation could compress further. At 18.7x current-year non-GAAP P/E, INTU is cheap relative to growth — but it's not so cheap that the market can't keep pushing it lower on narrative alone. The SaaS AI disruption narrative hasn't fully resolved.

The Math

Factor decomposition (trailing 250d):

R² = 22.5%. Idio variance 77.5% clears the 75% threshold — returns are stock-specific, not closet-index.

Valuation:

• FY2026 non-GAAP EPS guidance: $22.98-$23.18 → P/E 18.7x
• NTM P/E (blending FY2027): ≈16.3x
• Revenue growth: 17-18% H1, guided 12-13% full year
• GAAP operating margin expanding: 15.0% → 18.4% (Q2 YoY)
• OCF FY2025: $6.2B. FCF yield ≈5%.

Capital return: $1.8B buyback H1 (+40% YoY), $3.5B remaining authorization, dividend +15% YoY. Total capital return running at ≈6% of market cap annualized.

Decision

KEEP. 75% confidence INTU matches or outperforms QQQ over 15 weeks.

The trailing alpha is a rearview mirror. The income statement, the insider signal, the tax season data, and the valuation floor all point the same direction. Removing a 17% revenue grower at 16x NTM earnings with every executive cancelling their selling plans — while sitting at RSI 31 with the company's biggest seasonal catalyst 8 weeks away — is textbook negative timing skill.

The bear case is momentum continuation and narrative. The bull case is fundamentals, capital return, and a specific catalyst. For a 15-week window, the catalyst wins.

Monitor: May 21 Q3 earnings. If TurboTax revenue growth decelerates below +8% or customer growth turns negative, reassess. Watch QBO Capital charge-off rate — above 12% annualized would flag macro stress in the SMB base.