ServiceNow AI Business Tops $1 Billion in Contract Value as Agents Scale

ServiceNow said on July 22 that its AI products have passed $1 billion in annual contract value for the first time, a milestone that gives enterprise buyers one of the clearest signals yet that agentic AI is moving from pilots into paid, production workflows. The company beat Wall Street estimates for the second quarter and raised its full-year subscription outlook, according to its earnings release filed with the SEC.

For CIOs, the numbers matter less as a stock story and more as evidence of how large enterprises are actually buying AI: inside existing workflow platforms, with governance attached, and increasingly on usage-based terms.

What did ServiceNow report for Q2 2026?

Subscription revenue reached $3.877 billion, up 24.5% from a year earlier, or 23% in constant currency. Total revenue was $3.987 billion, up 24%. Current remaining performance obligations, the contracted revenue ServiceNow expects to recognize over the next 12 months, rose 21% to $13.2 billion. The non-GAAP operating margin was 29.5%.

Large customers kept growing. ServiceNow ended the quarter with 658 customers paying more than $5 million a year in contract value, up about 23%, and closed 123 deals worth more than $1 million in net new contract value, up nearly 40%.

The company raised its 2026 subscription revenue outlook to between $15.76 billion and $15.78 billion, about 22.5% growth. For the third quarter it guided to subscription revenue of roughly $3.98 billion, up 20.5%. CEO Bill McDermott called ServiceNow "a defining company that is only just getting started" in the release.

How fast is agentic AI growing at ServiceNow?

McDermott said agentic deployments grew ninefold over the past nine months, SiliconANGLE reported. CFO Gina Mastantuono said net new AI bookings again beat internal expectations and credited AI Control Tower, the company's product for discovering, monitoring and governing AI systems, with lifting its security and risk business.

On the earnings call, executives said net new AI contract value grew more than 40% from the prior quarter and first-time AI buyers rose more than 45% year over year, according to a GuruFocus summary of the call. Management also said more than 40 customers using an AI-driven Level 1 IT service product resolve 80% to 85% of service requests without a person stepping in, and that AI is tracking ahead of a goal of 30% of contract value by 2030.

That last data point is the one practitioners should test against their own experience. Resolution rates in that range typically depend on clean service catalogs, accurate configuration data and well-defined escalation rules, not just the model. Speakers at AI Connect events have made the same point repeatedly in discussions of what it takes to move agents from human-in-the-loop to agent ops.

Why does AI governance show up in the revenue story?

The Control Tower comment is worth noting. ServiceNow is selling governance as a product, not just a feature, and says it is pulling through security and risk revenue. That fits a pattern AI Connect has heard from enterprise teams: oversight that is designed in early lets companies deploy more, not less.

IBM's Ryan Zombo, Director of Data & AI Technical Sales, made the case with an analogy at COLLIDE 2025: "You put brakes on a car so that you can go fast." He spoke in a session on the five questions enterprises must answer to scale agentic AI. One of his questions was how to scale agents across business units without agent sprawl, which is precisely the problem a control tower product is meant to address.

What should enterprise buyers watch?

Three details deserve attention before renewals.

Pricing is shifting. Executives described a hybrid model that combines licenses with usage tied to outcomes, and said price uplifts for AI-native product tiers follow a 20% to 30% framework, per the GuruFocus summary. Buyers should model what happens to cost if agent usage grows as fast as ServiceNow expects.

Some of the beat was timing. ServiceNow said strong U.S. federal demand pulled some on-premises subscription revenue forward from the third quarter into the second, which partly explains why third-quarter growth guidance is lower.

Margins face pressure. Management flagged near-term gross margin pressure from hyperscaler partnerships and AI consumption. Vendors under margin pressure often revisit packaging, so contract terms on AI usage are worth locking down now.

Investors had been skeptical going in. The stock had fallen about 35% this year before the report, and shares rose more than 3% in late trading after the results, SiliconANGLE reported.

The bigger signal for enterprise leaders is the shape of demand. A billion dollars in AI contract value at one workflow vendor suggests that agent spending is consolidating on platforms that already hold the process data and the permissions model. The open question for the second half is whether customers see the outcomes to justify the usage bills, which is the measure AI Connect speakers have urged teams to set before they scale.

Frequently asked questions

How much AI revenue does ServiceNow have?

ServiceNow said its AI products surpassed $1 billion in annual contract value in the second quarter of 2026, the first time it has crossed that mark.

What is ServiceNow AI Control Tower?

It is ServiceNow's product for discovering, monitoring and governing AI systems and agents. The CFO said it is helping drive growth in the company's security and risk business.

How is ServiceNow pricing AI?

Executives described a hybrid model that combines licenses with usage tied to outcomes, with price uplifts for AI-native tiers following a 20% to 30% framework.

Filed under: Agentic AI, AI Strategy & ROI, AI Governance & Trust