Q: ServiceNow’s stock is down nearly 50% from its 2025 high. What happened?

A: A single fear took over the market: that AI agents will do the work people used to do inside platforms like ServiceNow, shrinking the number of paid “seats” companies need. Add a premium valuation with no room for disappointment, and the stock became an easy target once sentiment turned.

Q: Is that fear actually showing up in the numbers?

A: Not yet. Subscription revenue grew 22% last quarter. The renewal rate held near 97%. Contracted revenue not yet billed, the best gauge of forward demand, grew even faster than reported revenue, up over 22%. None of that looks like a business losing customers to AI. If anything, ServiceNow’s AI product is its fastest-growing line, and management just raised its 2026 target for it by 50% in a single quarter.

Q: What’s the honest weak spot, then?

A: Margins, a little. Subscription gross margin slipped from about 84.5% to 81.5% over the past year, the cost of building AI features and digesting three acquisitions at once. It’s a real yellow flag worth tracking, not a crisis. Operating margin still sits around 32%, and the company converts roughly a third of every revenue dollar into free cash.

Q: What would prove the bears right?

A: A handful of clean signals: the renewal rate slipping below 95%, the AI product’s growth decelerating instead of accelerating, or subscription growth falling into the low teens because customers are genuinely cutting seats, not just delaying deals. None of that has happened. If two of those show up together over the next few quarters, the thesis breaks.

Q: So, buy the dip or wait it out?

A: The valuation makes the decision easier than the headlines suggest. A year ago this stock traded near 60 times forward earnings. Today it’s about 25 times, for a business still growing over 20% with a fortress balance sheet and no debt problem. That’s a real repricing, not a rounding error. The honest case is a moderate starter position now, added to on further weakness, with earnings on July 22 as the next real test. It’s a Buy, not a maximum-conviction bet: the seat-based risk is genuine, just not confirmed yet.

The full scorecard, valuation methodology, and scenario modeling are in the complete ServiceNow (NOW) report on our Reports page.