ServiceNow ($NOW): The Governance Bet Behind 2026's Software Selloff

Some stocks fall because the business is breaking. Others fall because the market has decided, almost overnight, to reprice an entire industry's future. ServiceNow (NYSE: NOW) is squarely the second kind and that gap between the price action and the underlying numbers is what makes it worth a proper look.

The paradox in one paragraph

At its peak last summer, ServiceNow traded above $211 a share (split-adjusted, the company did a 5-for-1 split in December 2025). By early June 2026 it had fallen as low as $81, a drawdown of roughly 60% at the trough, for a company that beat guidance on every metric in its most recent quarter, raised its full-year outlook and still carries a 97% customer renewal rate. That combination, a beaten-down chart sitting on top of an accelerating business, is the reason this one is on our radar.

Why the stock actually fell

The trigger was not company-specific. Starting in late January 2026, a fear swept through enterprise software: that AI agents would do the work employees currently do inside these platforms, shrinking the number of paid seats companies need to buy. ServiceNow largely prices by seat, so the logic went that agentic AI was a direct threat to its business model. The whole sector got repriced on that fear, not just ServiceNow. Salesforce, Workday, Adobe, HubSpot and Oracle all fell alongside it through the spring.

The moment that crystallised it was ServiceNow's Q1 2026 earnings call on April 22. The quarter itself was clean: subscription revenue of $3.67 billion, up 22% year over year, current remaining performance obligations up 22.5% and full-year guidance raised. The stock fell roughly 18% the next day anyway, its worst single session on record. Management flagged delayed Middle East deals and margin pressure from the newly closed Armis acquisition, but the bigger story was that the market was not repricing the next quarter. It was repricing the next decade, deciding a 20%-plus compounder might not compound the same way once AI agents are doing more of the actual work.

The case that the market has this backwards

The counter-argument, and the one we find more persuasive, rests on a distinction between two layers of enterprise software. Execution-layer tools do specific tasks: filing a ticket, drafting a document or answering a routine question. Those are genuinely exposed to AI agent substitution. Orchestration-layer platforms are different. They govern, audit and coordinate what happens across an organisation's entire technology stack, holding years of approval chains, compliance rules and institutional memory that would take a multi-year re-architecture to replace. ServiceNow is almost entirely the second kind of business and the more autonomous AI agents a company deploys, the more it needs something watching over what those agents are allowed to do.

Nvidia's Jensen Huang made a version of this case at ServiceNow's own conference in May and it's worth quoting his actual words rather than the social media paraphrase circulating since: "for the first time service is software and the service industry is 100 times larger than the software industry." That is a claim about the addressable market expanding, not a stock price prediction and the distinction matters given how much distorted commentary about this stock is floating around X right now. Treat any post claiming Huang predicted the stock itself could "100x" as a misreading at best.

The numbers back the governance thesis reasonably well so far. The renewal rate has held at 97 to 98% for five straight quarters. Customers spending over $5 million a year with ServiceNow grew from 516 to 630 in twelve months. Roughly half of net new business now comes from consumption-based pricing rather than per-seat licences, which structurally weakens the "fewer seats means less revenue" bear case. And Now Assist, the AI product line the market fears will cannibalise the rest of the business, is instead the fastest-growing part of it: management raised its 2026 AI contract value target from $1 billion to $1.5 billion mid-year.

Insiders are buying too

For what it's worth, and insider activity is a supporting signal rather than a thesis on its own, ServiceNow's leadership has put real money behind this view. In February 2026, five senior executives, including CEO Bill McDermott, CFO Gina Mastantuono, Vice Chairman Nicholas Tzitzon, Chief People and AI Enablement Officer Jacqueline Canney and Special Counsel Russell Elmer, cancelled their pre-arranged Rule 10b5-1 selling plans, halting all scheduled future sales. McDermott separately committed to personally buying $3 million of stock, at the earliest date SEC short-swing rules allowed him to. Worth noting that the timing was partly rule-constrained rather than pure market-calling skill, but cancelling planned sales entirely is a deliberate act that goes beyond a routine buyback headline.

McDermott has also tied his own compensation to an explicit target: a $1 trillion market cap by 2030, up from roughly $109 billion today. That's a bold, self-serving number to put in public and we wouldn't treat a CEO's own prediction as objective evidence. But it does signal how much conviction management is putting behind the growth plan, which targets subscription revenue doubling from around $15 billion to $30 billion by decade's end.

What still needs to go right

None of this makes ServiceNow a clean story. Growth is decelerating from the high-20s percentages of a few years ago toward the low-20s now, still strong but a genuine trend worth watching rather than dismissing. The company is digesting three acquisitions at once, Armis, Moveworks and Veza and integration risk on that scale is real even without the AI narrative attached. The Middle East deal delays flagged in Q1 could recur. And while the orchestration-layer thesis is coherent, it's a thesis, not a certainty. If AI agent capability advances faster than ServiceNow's governance tooling can keep pace, the competitive dynamics could shift in ways that are hard to model from here.

On valuation, the answer depends on which lens you use

This is a case where the standard shorthand metrics genuinely disagree with each other, so it's worth being specific rather than quoting a single number. Trailing P/E sits around 60 to 65x, which looks expensive, but that denominator is distorted by GAAP earnings squeezed by acquisition costs and stock-based compensation this year. Forward P/E is closer to 24 to 25x. Price to sales sits around 7x, down from the 8x-plus it commanded before the selloff. On forward free cash flow, using management's guided margin, the multiple works out somewhere in the high teens.

PEG ratio is the clearest example of why methodology matters here. Using trailing twelve-month EPS growth, artificially suppressed by one-off costs, PEG comes out around 4.5, which looks stretched. Using forward-looking five-year expected earnings growth instead, the same stock prices out to a PEG close to 1.0, Peter Lynch's textbook definition of fair value. Both numbers are technically correct. They're just answering different questions, and which one you trust depends on how much weight you put on this year's distorted GAAP earnings versus the underlying growth trajectory.

Wall Street's own numbers land on the optimistic side. Analyst consensus sits at Strong Buy, with an average twelve-month price target around $140, roughly 30% above where the stock trades today.

What we're watching next

Q2 2026 results are due after market close on July 22. Two numbers will tell us the most: constant-currency current remaining performance obligations, where a reading above 21% keeps the growth story intact and a drop toward 19% would be a real yellow flag, and the renewal rate, which needs to hold at 97 to 98%. A clean beat on both, alongside continued Now Assist momentum, would meaningfully strengthen the governance-layer thesis. A miss on either, particularly if it comes with commentary about seat count pressure rather than just deal timing, would be the first real data point supporting the bear case rather than just the narrative.

The bottom line

ServiceNow isn't a stock we'd call cheap in any absolute sense, but it's priced as though its next decade looks meaningfully worse than its last two, while the actual numbers coming out of the business so far say otherwise. The AI disruption fear that took over 50% off the share price was a real risk worth pricing and it may yet prove partly correct. But the specific mechanism the market feared, agents replacing seats, doesn't show up in the renewal rates, the customer expansion data or the AI revenue line, all of which are moving the opposite direction. That's the kind of gap between narrative and numbers worth tracking closely into the next print.

This article is for informational and educational purposes only and does not constitute financial advice. Always do your own due diligence before making investment decisions. The author may hold positions in securities mentioned.