Intuit (INTU) Q4 FY2026: 3% Online Customer Growth Matters More Than the EPS Guide

The headline that moved the stock was a fiscal 2027 non-GAAP EPS guide of $22.88 to $23.12 against a consensus of $27.30, and the headline was wrong. Effective August 1, Intuit stopped excluding share-based compensation from its non-GAAP measures. The guide carries a $5.81 per share SBC charge that consensus did not. Add it back and the like-for-like range is $28.69 to $28.93, roughly five percent above the street. The first quarter is the same arithmetic in the other direction: $2.44 to $2.48 plus $1.48 gives $3.92 to $3.96 against $4.02, a shortfall of about two percent. Full year above, opening quarter slightly below. That is a spending phasing issue, not an earnings collapse, and it is not why the shares fell 7.3 percent after the close.

They fell because revenue guidance came in at $23.279 billion to $23.512 billion, growth of nine to ten percent against the fourteen percent just delivered, and because of what sits underneath the deceleration. TurboTax is guided to two to three percent. Credit Karma decelerates to eleven to thirteen from twenty. Mailchimp is guided to negative one to zero and is being carved into its own reportable segment beginning in fiscal 2027, which is how a company sets up either a fix with a scoreboard attached or an exit. Global Business Solutions at thirteen to fourteen is the only line holding, and it is holding on ARPC rather than units.

The unit disclosure has been telling this story for a year and nobody read it. Total US TurboTax federal units fell to 39.0 million from 39.9 million, down two percent, with online units down two percent and desktop down seven. TurboTax revenue grew seven percent on that base. All of it was price. Management has now said out loud that price is the number one reason customers leave TurboTax and that the company lost quality DIY filers to lower-cost providers. That admission is a restatement of a table Intuit published in its own press release four quarters running. The company optimized a shrinking unit base for ARPC until the ARPC stopped covering the attrition, and the point at which those two curves crossed was fiscal 2026.

Total online paying customers grew three percent to 8.9 million, roughly two points slower than the prior year. That is the number the guidance is built around and the number the entire fiscal 2027 plan is designed to fix. QuickBooks Free and QuickBooks Lite have taken more than 20,000 customers in their first month. Intuit is accepting lower initial ARPC in DIY tax to rebuild the funnel. A business with intact pricing power does not launch a free tier, and this is the correct response to not having one.

What funds it is the seventeen percent workforce reduction taken in May. GAAP operating income is guided to $7.408 billion to $7.490 billion, growth of twenty-six to twenty-seven percent on nine to ten percent revenue, which puts GAAP operating margin near 31.8 percent against 27.4 percent in fiscal 2026. Some of that is the absence of the $293 million restructuring charge booked in the fourth quarter. Most of it is headcount that no longer exists. Fiscal 2027 is therefore a single trade stated plainly: buy customers with price, pay for the price cuts with people, and hold margin expansion anyway.

The cash flow statement contains the finding the release does not flag. Operating cash flow rose to $8.838 billion from $6.207 billion, up forty-two percent. Of that $2.631 billion increase, $1.714 billion is the deferred income tax line swinging from negative $435 million to positive $1.279 million. Long-term deferred income tax assets on the balance sheet fell to $172 million from $1.222 billion. The effective rate moved to approximately twenty-four percent from twenty. The shield is substantially consumed, and cash taxes rise from here independent of how the operating year goes. Sixty-five percent of the headline cash flow improvement was a non-cash reversal that does not repeat.

The moat question is where the market and the business disagree most. What AI plausibly commoditizes is the eighty-dollar self-serve tier — a rules engine wrapped in a wizard, which FreeTaxUSA and Cash App Taxes were already commoditizing before anyone had an agent. What it does not touch is IRS-authorized e-file transmission, the bank and payroll integrations, the accuracy-and-audit liability Intuit underwrites, and the accountant channel that sells QuickBooks Desktop Enterprise into mid-market firms. Those are licensing and distribution assets, not model-capability assets, and an agent that files a return still has to hand it to someone holding an EFIN. Meanwhile the AI-native side is the part working: TurboTax Live revenue grew thirty-seven percent with customers up thirty-eight, the Big Bets collectively grew thirty-four percent and reached thirty percent of full-year revenue, and QuickBooks Online Accounting grew twenty-three. The done-for-you business compounds while the self-serve business gets commoditized underneath it. The bear case is not that AI kills Intuit. It is that the mix shift arrives faster than the assisted business can absorb it, and the January-through-April season is when that gets settled.

Capital return remains the cleanest part of the print. Intuit repurchased $5.5 billion in fiscal 2026, up ninety-six percent, cutting weighted-average diluted shares two percent net of a $2.056 billion SBC drag, with $7.9 billion of authorization remaining. The dividend rose fifteen percent to $1.38. Cash and investments of $7.2 billion against $7.7 billion of debt leaves the balance sheet roughly neutral, with $1.75 billion of June senior notes covering fiscal 2027 maturities and $1.249 billion sitting in short-term debt. The company also retired $27.006 billion of treasury stock, a housekeeping entry that explains the decline in stockholders’ equity to $18.992 billion.

At roughly $357 after hours the shares are down about forty-three percent year to date and forty-seven percent from the peak, a market capitalization near $97 billion, or 17.6 times the midpoint of GAAP fiscal 2027 guidance and about 12.4 times the SBC-adjusted non-GAAP equivalent. Street average target into the print was $390.72 across eleven buys, six holds and two sells, and that average gets cut this week. A securities class action carries a September 8 lead plaintiff deadline. Base case is $340 to $400, seventeen to twenty times GAAP, on a year that executes the reset without reaccelerating anything. Bull case is $470 to $500 if QuickBooks Free converts at a rate that turns online paying customer growth back above six percent and the cheaper DIY entry point turns federal units positive in the coming season. Bear case is cohort derating to $250 to $280, twelve to fourteen times GAAP, alongside Salesforce, ServiceNow, Adobe and Atlassian as the market applies one AI-disruption multiple to every seat-and-subscription franchise regardless of whether the underlying attrition is a model problem or a price problem — the 52-week low of $278.94 set on July 23 sits inside that range. The September 17 Investor Day is where management has to show the funnel work in numbers rather than intent.

The decision-relevant number is not the guide, the margin, or the buyback. It is the TurboTax federal unit table in the fiscal third quarter release next April. Intuit is giving up ARPC to buy units. If total US units do not turn from negative two percent to positive, the price concession bought nothing and the deceleration is structural.