Forty-five percent. That’s how far ServiceNow has fallen from its highs. Not because the business deteriorated. Not because customers started leaving. Not because the model stopped working. The stock is down nearly half because the market decided to price in a future that hasn’t arrived yet — and may not.

Here’s what actually happened last Wednesday. ServiceNow reports Q1 2026 results. Beats the high end of every single metric it guided for. Raises full-year subscription revenue guidance. Increases its AI annual contract value target from $1 billion to at least $1.5 billion. And the CEO — not the company, the CEO personally — bought $3 million of stock with his own money back in February, before any of this played out publicly. The stock dropped 17% the same day results hit.

That’s not a reaction to bad news. That’s a repricing of fear.

The specific fear is this: AI agents are going to eliminate enterprise headcount, enterprise headcount is what drives SaaS seat licenses, and therefore the entire ServiceNow revenue model is sitting on a foundation that’s about to crack. The market looked at Anthropic’s expanding Claude agent ecosystem and connected a straight line to ServiceNow’s customer count. And then it sold first and asked questions later.

Subscription revenues came in at $3.67 billion for Q1 — up 22% year over year. Total revenues hit $3.77 billion. Current remaining performance obligations, which captures contracted revenue locked in over the next 12 months, grew 22.5% to $12.64 billion — a 100 basis point beat. Total remaining performance obligations reached $27.7 billion, up 25% year over year. The long-term revenue base is actually growing faster than the near-term one. That’s not a red flag. That’s the opposite.

Customers spending more than $1 million in Now Assist ACV grew over 130% year over year. Large deals — transactions over $5 million in new ACV — jumped nearly 80% with 16 closing in the quarter. New logo ACV growth came in above 50% year over year, including what management called the largest net new logo deal in company history at over $15 million. Renewal rates held at 97%. Non-GAAP operating margin landed at 32%, 50 basis points above guidance. Free cash flow margin was 44% for the quarter.

The company also executed a $2 billion accelerated share repurchase in Q1 — buying back roughly 20.1 million shares, double the total repurchased in all of 2025. Another $4.2 billion sits under the current repurchase authorization.

None of this looks like structural decline.

What Most People Are Glazing Over

The seat-license concern is real. ServiceNow knows it. They’ve been restructuring the pricing model for over a year. As of Q1 2026, 50% of net new business comes from non-seat-based mechanisms — tokens, infrastructure usage, connectors, consumption-based contracts. The CEO has addressed this directly on every call since Q3 2025. The CFO added on the Q1 call that AI investment at large enterprises is being funded from three places: new budgets, yes, but also labor cost reductions and consolidation of point solutions. ServiceNow benefits from all three. That’s the part the fear trade is ignoring.

Slight tangent, but worth saying: the platform has processed 95 billion workflows and more than 7 trillion transactions. That’s 22 years of enterprise operational data baked into the system. An AI model that shows up with no workflow history, no process context, no organizational memory — that’s a tool. ServiceNow is positioning as the layer where AI actually does work inside an enterprise, not just generates output. That distinction matters more than the market is currently giving it credit for.

Multi-product AI adoption is also accelerating in a way that gets lost in the headline numbers. Deals including three or more Now Assist products grew nearly 70% year over year in Q1. Thirty-six deals closed with five or more products. The average deal size for the AI Control Tower product more than doubled quarter over quarter. This isn’t one AI feature getting bolted onto an existing contract. It’s a company expanding its footprint across the enterprise at the exact moment the market thinks the model is getting disrupted.

Two acquisitions are weighing on near-term margins. The Armis deal — $7.75 billion in cybersecurity, closed ahead of schedule — is pressuring full-year operating margin to a guided 31.5% and free cash flow margin to 35% for 2026. Full normalization is management’s 2027 call. On top of that, the Veza acquisition closed in March, adding access visibility capabilities for AI agents and enterprise data. Neither acquisition contributes meaningful revenue in the near term. Both add cost ahead of scale. That’s the honest explanation for why margins look compressed right now — and it’s a deliberate investment decision, not evidence of a broken model.

The Valuation After the Damage

NOW now trades at roughly 5.5x forward price-to-sales and a forward P/E near 20x. For a company compounding revenue at 22% annually, holding a 97% renewal rate, generating $4.6 billion in trailing free cash flow, and sitting on $27.7 billion in total remaining performance obligations. Full-year subscription revenue guidance was raised to $15.74–$15.78 billion — roughly 22–22.5% growth for the year.

What triggered the selloff wasn’t bad guidance. It was flat organic full-year guidance against a market that wanted acceleration. Management also deferred the full consumption revenue conversation to the May Financial Analyst Day rather than spelling it out on the earnings call. That combination — no upside surprise on the guide, no immediate answer on consumption trajectory — is what caused the 17% move. The business didn’t change in a day. The market’s patience for the AI monetization arc did.

• Bull case: AI Control Tower becomes the default enterprise AI orchestration layer. Now Assist ACV reaches $2 billion or more by 2027. Armis integrates cleanly and margin recovery is faster than expected. New logo momentum holds above 50%. The stock works back toward 30x forward earnings over a 2–3 year hold.
• Base case: Growth stays in the high teens to low twenties. Margins improve gradually as acquisition costs fade. Buybacks support the valuation floor. Mid-to-high teens annual returns from current levels over a multi-year hold.
• Bear case: Agentic AI accelerates faster than the pricing pivot can absorb. Armis integration disappoints. Middle East deal delays become recurring rather than one-time. Growth decelerates into low double digits. The multiple contracts further from an already compressed level.

Knowledge 2026 runs May 5–7 in Las Vegas. Conference attendance is up 11% year over year. The Financial Analyst Day scheduled around the event is where management is expected to lay out the consumption revenue trajectory in full — the answer the market wanted on the earnings call but didn’t get. Three things to watch: Now Assist ACV momentum, the Armis integration cost update, and whether the organic full-year guide gets any upward movement. Those three data points will do more to define the next move than anything else between now and Q2 earnings.

The business is growing at 22%. The stock is priced for something closer to a slow bleed. Either the fear resolves and the gap closes — or the data starts confirming what the market suspects. Right now, the numbers aren’t confirming it. Eight days until Las Vegas.