ServiceNow: The Market Still Prices the AI Scare. The Cash Flow Says Otherwise.

ServiceNow: The Market Still Prices the AI Scare. The Cash Flow Says Otherwise.

The market is still pricing ServiceNow as the AI-scare casualty it stopped being months ago. The shares fell more than 40% in the first four months of 2026 on the fear that the workflow-software maker would be flattened by the exact AI models it now sells into its own platform. May produced the best monthly performance since available data going back to 2012, with the stock up more than 40% in a single month as AI fears faded across enterprise software. And the recovery still has not finished: entering this week the stock was down roughly 17% year to date and about 28% over the trailing twelve months, roughly 35% below its 52-week high of $194.73, even after a tape that has climbed more than 30% in the past month and added about 6% in the latest session. The panic ended. The repricing of the company has not.

Everything turns on whether the business kept working underneath while the tape was screaming. It did not just keep working; it accelerated. In the quarter ended June 30, ServiceNow beat the high end of its own guidance on every topline and profitability metric. Subscription revenue, the engine of the model, rose 24.5% year over year to $3.877 billion, and total revenue grew 24% to $3.99 billion. Non-GAAP earnings of 90 cents came in ahead of the roughly 86-cent consensus, up 11% from a year earlier. The forward-looking number investors watch hardest is current remaining performance obligations, or cRPO — contracted revenue the company must still deliver over the next twelve months, the closest public proxy for booked demand. It reached $13.20 billion, up 21% year over year.

Here is the part that makes the old story stale: the headline still reads as decelerating, while the demand metrics run ahead of it. Management guided the next quarter's subscription growth down to about 20% on a constant-currency basis, but the leading indicators never slowed down. ServiceNow signed 123 transactions over $1 million in net new annual contract value in the second quarter, up nearly 40% year over year, and ended the period with 658 customers spending more than $5 million a year. ServiceNow AI crossed $1 billion in annual contract value, and the CEO said agentic deployments — autonomous AI workers running on the platform — increased ninefold in nine months. At the May analyst day the company laid out a path to more than $30 billion of subscription revenue with AI eventually making up 30% of everything it sells. The crowd watches the guided 20% headline. The booked business says the next twelve months are faster, not slower.

That gap between the guide and the bookings is the setup, and it is exactly why an October earnings report is a low hurdle. When the second quarter printed, management guided third-quarter subscription revenue to $3.975–$3.980 billion — roughly 20% constant-currency growth, stripping out exchange-rate swings, and a visible step down from the 24.5% just reported. Investor's Business Daily's blunt summary was "sales outlook light," and the raised full-year guide still got called conservative. That is the market-anchoring moment.

The thing is, the third-quarter bar is lower than the demand underneath it, for mechanical reasons. ServiceNow pulled some federal on-premise subscription revenue out of the third quarter into the second, and a stronger dollar is worth about a $35 million headwind to third-quarter cRPO growth on its own. Take those two out and the guided 20% understates what the quarter is actually running at. The first quarter did the same thing — the company said it beat the high end of guidance across all Q1 2026 topline growth and profitability metrics — and then guided carefully again. A management team that beats the high end of its own guidance in back-to-back quarters and then sets a mechanically depressed bar for the third is a team that keeps giving buyers a low hurdle to clear. A quarter of federal timing and a currency swing are not the kind of thing that changes a cash-flow engine.

That engine is the real reason to care, because the single print is not the thesis; what the print does to next year's numbers is. The financial bridge is free cash flow, and it is already visible. Trailing free cash flow is $4.57 billion, up 20% year over year at a 31% margin. The second quarter's $634 million looks thin only because this company books the bulk of its billings — and therefore its cash — in the back half, as it always does. For the full year, management raised subscription guidance to about $15.77 billion, implied 21% constant-currency growth, held operating-margin guidance at 31.5%, and kept a 35% free-cash-flow margin on the books. The cash engine is getting bigger, not smaller.

Now put that against the price. At a roughly $131.5 billion market cap, the stock trades near 23 times projected 2026 free cash flow of about $5.7 billion — the 35% guidance margin applied to roughly $16 billion of expected revenue. Measured on trailing free cash flow it is closer to 29 times, and at its early-year peak that same fast-growing cash stream carried roughly 50 times. The multiple has roughly halved while the cash flow grew about 20%. That is the definition of a reset that has not yet repriced. On an income-statement basis the shares will never look cheap — they still go for about 78 times forward earnings — which is why the free-cash-flow lens is the honest one for a business whose reported profit is understated by heavy stock-based compensation and depreciation.

Carry the bridge into next year. If subscription growth simply holds in the low 20s and the free-cash-flow margin drifts toward the mid-30s as the operating-margin guidance implies, 2027 free cash flow lands near $7 billion. A return to something like 26 times that — comfortably below the multiple the stock carried at its peak, but above the reset multiple of roughly 23 times 2026 cash flow today — puts fair value around $175, about 38% above the current price, over the next twelve months. That target is not a round trip to the euphoric high. It is asking the multiple to recover part of what the scare took out while the cash flow keeps growing into it. AInvest's aggregate signal already labels the stock a Buy, with the fundamental inputs the strongest piece of the score. The debate over whether this business is the AI winner is largely settled at the aggregate level; the remaining fight is the multiple. That is a better fight to be on the right side of.

Two things could break this, and they deserve to be named plainly, because I can be wrong again. The first is gross margin. Management guided full-year gross margin down to 81% and pinned it on heavier hyperscaler usage and faster AI adoption — AI can become a less profitable way to sell software if the compute and infrastructure cost keeps creeping. The 35% free-cash-flow margin guide absorbs that for now through operating leverage, but a guide that keeps slipping would undermine the entire bridge. The second is demand, and it is the tripwire. If third-quarter cRPO growth lands below the 20% constant-currency guide, or if the street starts cutting rather than raising 2027 estimates, then the reset was wrong and the position does not belong. The position: own the stock into the October print, hold through whatever volatility the reaction brings, and cut on the tripwire without ego — the re-entry is available if the setup resets. The thesis is the cash-flow bridge, and the cash-flow bridge is doing exactly what it is supposed to do.