Here is the question worth spending fifteen minutes on this morning: when a software company beats on revenue, beats on cRPO, beats on operating margin, raises its full-year free cash flow guidance, and still loses roughly one dollar in six overnight, which part broke, the business, or the price?
ServiceNow reported Q3 results after the close on October 28, 2026. The company reported adjusted EPS of $1.00, below the $1.02 analysts expected, while revenue was about $4.11 billion, slightly ahead of the $4.09 billion forecast. Shares closed the regular session at $150.13, down 3%, and traded around $123 after hours, down about 18%.
But read past the EPS miss and the numbers look different. ServiceNow exceeded guidance across all key metrics: subscription revenue grew 20.5% in constant currency, cRPO grew 21% in constant currency, and operating margin was 32.5%, all ahead of internal targets. The company raised its full-year AI ACV target to $1.7 billion, up from previous guidance of $1.5 billion, reflecting accelerating monetization of artificial intelligence. It also raised full-year free cash flow margin guidance to 36%.
The EPS miss itself was narrow and explainable. The company’s interest income declined year-over-year, while interest expense rose after it took on debt to help fund its Armis acquisition. That is a balance sheet consequence of buying a security company, not evidence that the workflow business is softening.
The more substantive concern is Q4’s cRPO outlook. Fourth-quarter cRPO guidance came in at 18.5% in constant currency, reflecting a 25 basis point headwind from a higher mix of on-premises deals. That deceleration from Q3’s 21% spooked investors hunting for acceleration, not steadiness. Management acknowledged that strong early renewals in Q3 contributed about half of the cRPO beat, pulling deals forward. In other words, Q3 stole some of Q4’s credit. The company also did not raise full-year subscription revenue guidance by the full Q3 beat, which may have disappointed investors expecting a larger raise.
None of that changes what the underlying business is doing. AI net new ACV grew 65% year-over-year in Q3. Control Tower net new ACV grew over 6x year-over-year. The agentic product EmployeeWorks reached $100 million ACV in less than three quarters. The number of deals over $1 million in AI ACV grew 4.5x year-over-year. The company reported a 98% renewal rate and ended the quarter with 680 customers generating more than $5 million in ACV, including 31 additional customers surpassing $20 million in ACV over the past year.
The strategic context matters here. This is the first major enterprise software result since agentic AI moved from pilots into production. ServiceNow has said year-to-date consumption of agentic AI in production is up more than 13x year-over-year, and previously said agentic deployments increased ninefold over nine months. ServiceNow is not describing a future state, it is describing a customer base actively deploying agents and paying to govern them.
The bull case rests on that production reality. If agentic AI is genuinely arriving in enterprise workflows, ServiceNow is positioned as the governance layer that captures spend from every third-party agent running inside a customer’s environment, not just its own. On the earnings call, management described ChatGPT as an interface into ServiceNow, but said the underlying execution still consumes ServiceNow “assists”, preserving consumption-based monetization.
The bear case is valuation. ServiceNow had already shed roughly 3% during the regular October 28 session ahead of the report, as investors trimmed exposure to a stock with little room for error. Intensifying competition from Microsoft, Oracle, and Salesforce is a real headwind, as those players are investing significantly to deliver enterprise AI agents, workflow orchestration, governance, and security. A stock with any multiple expansion baked in cannot afford to miss EPS by two cents and guide Q4 cRPO below consensus without consequence.
Watch next: Whether management’s explanation for the Q4 cRPO step-down holds up in analyst follow-on conversations over the next week. If early renewals truly pulled forward Q3 demand, Q4 subscription revenue could still surprise. If the deceleration reflects genuine enterprise caution about committing to larger AI contracts, the Q1 2027 outlook becomes the real test. The AI ACV target raised to $1.7 billion gives the company a specific number to prove against. That is where the thesis either compounds or cracks.
