ServiceNow (NOW) Q2 2026 Update: The Target UBS Set, Blown Apart by NOW's Own Earnings

cRPO grew 21.5% year-over-year, beating guidance by 200 basis points; AI-native products carry a 20-30% pricing premium — the April downgrade thesis that "AI is eroding SaaS pricing power" just got overturned this quarter.

Chapter 1 | Earnings Checkup: Every Metric Cleared the High End of Guidance

CFO Gina Mastantuono's framing was direct: every revenue and profitability metric this quarter cleared the high end of guidance. This isn't a narrow beat — it's a beat across the board. That scorecard means more once you put it back in the context of April's debate.

Chapter 2 | The Target UBS Set on 4/10: Officially Blown Past This Quarter

On April 10, UBS downgraded NOW from Buy to Neutral and slashed its price target from $170 to $100 (-41%), citing "shaken confidence in enterprise software's positioning in the AI era" broadly, and cut its cRPO estimate to 16%. The stock gapped down sharply that day and touched an intraday low of $81.40, but was aggressively bought back before the close on roughly 2x average daily volume, closing with a long lower wick — PVL called that structure a "selling-exhaustion top" in the v2.1 article, betting UBS had it wrong.

Across two straight quarters, actual cRPO has landed in the 21-23% range, 5-6 percentage points above UBS's 16% assumption — that's not one lucky quarter, it's a systematic miss repeated twice in a row. UBS's downgrade logic wasn't wrong in principle (pricing power for AI-era application software is a real issue) — but its call to classify ServiceNow as "a generic application-software company that would get hurt" has been overturned by two consecutive quarters of results.

Chapter 3 | AI Pricing Power, Tested: Not Eroding — Rising

The most important new disclosure this quarter is management's first public figure for AI-native pricing: AI-native SKUs carry a 20-30% pricing premium over traditional products. The weight of that number only becomes clear against UBS's core logic — the April downgrade wasn't betting that "ServiceNow's growth would slow." It was betting that "AI gives customers more substitute options, shifting bargaining power from vendor to customer and structurally eroding pricing power."

If that thesis were correct, ServiceNow's AI products should need price cuts or bundling to move. What actually happened runs the opposite direction: customers aren't just willing to pay more for AI features, they're paying double-digit percentages more. CEO Bill McDermott also disclosed that agentic deployments grew 9x in nine months, that the company is closing in on a "Rule of 60" (revenue growth rate plus margin, combined), and that it's targeting AI at 30% of ACV by 2030.

Chapter 4 | AI Governance: The Real Source of Pricing Power Isn't the AI Features Themselves

The 20-30% pricing premium is the result, not the cause. The real question is: why are customers willing to pay more for it? The answer isn't in the AI features themselves — it's in AI governance, the topic management spent the most airtime on this quarter's earnings call, and the direction PVL believes CEO Bill McDermott should keep doubling down on.

How Big Is the Problem: 2.2 Billion New Identities Entering the Enterprise at Once

McDermott gave a staggering figure on the earnings call: "There are 2.2 billion agents entering the enterprise globally. That's 2.2 billion new identities." Every AI agent, once granted permission to execute tasks, becomes another autonomously acting identity inside the enterprise — one that can read data, trigger workflows, and operate across systems. Left ungoverned, these identities aren't a theoretical risk, they're an immediate one: "Every ungoverned asset and identity multiplies the blast radius of AI in the enterprise." Chief Product Officer Amit Zavery put it even more bluntly: the problem is "how do you find errors or issues with any of the rogue agents running around and stop them?"

Put plainly: a company that rushes to deploy AI agents for efficiency without simultaneously building the governance machinery — who can authorize an agent, who can audit it, who's accountable when something goes wrong, how to shut it off in real time — is effectively letting dozens or hundreds of invisible new employees roam freely through its systems, with no HR, no badge access, and no offboarding process. That's not walking a tightrope — it's running naked. For regulated industries (finance, healthcare, government), a single AI misfire with no audit trail can be a material compliance event; for an ordinary enterprise, an agent hijacked by prompt injection can become a straight-up security breach. A company that keeps scaling AI deployment without governance is planting a time bomb it can't predict the timing of — at that point, this isn't a risk-management shortfall anymore, it's a survival issue.

ServiceNow's Answer: AI Control Tower, 500+ Customers in Six Months

ServiceNow's headline product this quarter, AI Control Tower, was built exactly for this problem — it discovers, observes, governs, protects, and measures every AI deployment across the enterprise, and explicitly isn't limited to ServiceNow's own products. Zavery specifically emphasized: "It's not really for ServiceNow assets, but everything else." In other words, even when a customer is running third-party AI tools, ServiceNow wants to be the unified control layer sitting on top.

McDermott stated the positioning directly: "We're in the control business. One platform, one system of action, any cloud, any agent, any workflow, any model, governed, secured and accountable." Combined with its earlier acquisitions of Armis (real-time tracking of nearly 7 billion devices) and Veza (identity and access governance), ServiceNow is positioning itself as the "rules and rails" for enterprise AI deployment — not just a vendor supplying AI features, but the layer that decides which AI behaviors are allowed and which get blocked.

Governance and Pricing Power Are Two Sides of the Same Coin

Put this chapter's logic next to Chapter 3's 20-30% pricing premium, and the causal chain becomes clear: enterprises aren't paying more for "AI features" — they're paying more for the confidence to deploy AI at scale. CFO Gina Mastantuono put it directly: "Customers absolutely understand the value that our AI is delivering" — and a large part of that value is the sense of control governance provides. A company that hasn't first solved who's accountable when an agent misbehaves, how to audit it, and how to shut it off instantly hits a hard ceiling on AI deployment speed the moment it reaches compliance and security review — no matter how fast the underlying technology is. Conversely, whoever builds a solid governance layer first holds the key that lets customers deploy faster with confidence — and that key, not whether any single AI feature is good, is what actually keeps lifting pricing power over time.

Chapter 5 | Checking April's Tracking Thresholds: What Got Validated This Time

The v2.1 article set up a set of daily tracking metrics meant to test whether the 4/10 "selling-exhaustion top" call would hold up. Checking them against the latest numbers:

Three of the four metrics — share price, cRPO, and analyst price targets — have already clearly confirmed in the bull-case direction; the stock's position within the 52-week range is still in the middle, meaning the market hasn't yet fully priced in this confirmation — that's the gap still waiting to close, not a flaw in the thesis itself.

Chapter 6 | Valuation: EV/FCF Still at Half the Ten-Year Median

EV/FCF sitting at just half its ten-year median is the single number most worth remembering from this update: the market is still pricing ServiceNow as if "will AI erode pricing power" remains an open question, but this quarter's earnings have already answered it. If cRPO and the pricing premium keep confirming over the next one to two quarters, there's no reason for this valuation discount to persist.

Chapter 7 | Key Risks and Thesis-Breaking Conditions

1. Whether cRPO can hold above 20%: two straight quarters in the 21-23% range is a strong signal, but a third-quarter confirmation is still needed to rule out the possibility that one large single contract is propping up the number.
2. Durability of the AI pricing premium: the 20-30% premium was disclosed for the first time this quarter, and it remains to be seen whether it holds up in subsequent quarters or gets compressed by competition as AI feature penetration rises.
3. Whether the AI-governance thesis survives a real-world test: right now it rests on management's narrative and early adoption figures (500+ customers), and hasn't yet been stress-tested by a major security incident or regulatory audit — if a widely reported AI-agent governance failure occurs at a company that didn't choose ServiceNow's governance solution, the persuasiveness of this thesis would be tested.
4. Budget pressure on non-AI application software: the macro backdrop UBS's report flagged — budget pressure on non-AI application software — hasn't gone away, and could still affect renewal and expansion pace among marginal customers.
5. How fast the valuation discount closes: if the market keeps failing to correct the EV/FCF discount, that would suggest lingering doubt about the "pricing power has been validated" call, and would warrant re-examining whether the market is seeing a risk PVL hasn't.

Chapter 8 | Tracking Checklist and Conclusion

• Whether Q3 FY26 cRPO holds above 20%, a third confirmation of whether UBS's thesis has been overturned.
• Whether the AI-native pricing premium (20-30%) gets re-disclosed and holds up in next quarter's earnings.
• Whether AI Control Tower customer count keeps growing at the pace of its first-six-months 500+, and whether the company starts disclosing retention and expansion figures.
• Whether the stock keeps moving out of the middle of its 52-week range as earnings keep confirming, in the direction of closing the valuation discount.
• Whether agentic AI deployment growth (9x in nine months) holds up, and whether new customer-adoption cases emerge.
• Whether budget pressure on non-AI application software shows up among marginal customers.

Sources