ServiceNow Earnings (July 22): Subscription Revenue Up 24.5%, Guidance Raised, and the AI-Eats-SaaS Verdict Went Its Way

---

The Trial Returned a Verdict, and It Was Not IBM's

UPDATE (August 7, 2026): ServiceNow reported on July 22 and separated itself from the contagion this page put it on trial for. Subscription revenues of $3,877 million rose 24.5%, clearing the high end of the company's own guidance range by 150 basis points on a combination of net-new ACV outperformance and a richer on-premise revenue mix. US Federal demand and AI and cybersecurity spend were named as the drivers.

Guidance went up. Full-year 2026 subscription revenue guidance was raised to $15,760-15,780 million, roughly 22.5% growth. Q3 subscription revenue was guided to $3,975-3,980 million, about 20.5% growth and 20% in constant currency, with cRPO growth of 19.5%.

The AI line stopped being a slide and became a number. AI annual contract value passed $1 billion, with agentic AI deployments up 9x in nine months and AI Control Tower adoption above 500 customers in six months. That is the specific evidence this page said July 22 would either produce or fail to.

The counterweight: coverage of the results deck flagged margins under pressure from AI investment. A company can win the AI-eats-SaaS argument on demand and still pay for it on the operating line, and those are separate questions with separate answers.

Not sourced: a verified realised move against the ±8% implied quoted below, so the volatility framing on this page is left ungraded rather than guessed.

---

TL;DR

• ServiceNow reports Q2 results expected Wednesday, July 22, after the close, sharing the evening with Tesla.
• The context is brutal: IBM's software miss just produced its worst day since 1987 and NOW fell 7.5% in pure sympathy without reporting anything.
• That makes July 22 the first actual evidence in the AI-eats-SaaS trial: the sector's premium execution story either separates itself from IBM or confirms the contagion.
• Options price ±8%, elevated for NOW, and this quarter it's earned. The burden of proof, the tells, and the trade below.

Guilty by Association

NOW lost 7.5% on IBM's numbers. On July 22 it reports its own.

ServiceNow spent years as enterprise software's teacher's pet: the cleanest growth, the stickiest workflows, the premium multiple that never needed defending. Then IBM blamed a $660 million revenue miss on software weakness, and the market marked NOW down 7.5% inside an afternoon without waiting for evidence. That's what a sector on trial looks like: individual innocence stops mattering until someone proves the charge wrong with numbers.

NOW reports first among the pure-plays. Salesforce, Adobe and the rest of the cohort will trade off this print before their own. The stakes are sector-wide.

What Acquittal Requires

The bar changed shape after IBM. A standard beat-and-raise no longer clears it; the market needs specific counter-evidence to the AI-commoditization charge:

• Subscription revenue and cRPO growth holding rate. The thesis says AI erodes seat-based demand. Steady bookings growth is the direct rebuttal, and even one point of deceleration gets read as the IBM disease in its early stage.
• AI products as a named revenue driver, with numbers attached. "Strong customer interest in our AI capabilities" is 2025 language and the market will sell it. What acquits: disclosed AI-product contribution, attach rates, pricing uplift. Receipts, per the season's one rule.
• No "elongated sales cycles." After IBM, that phrase is a confession. Any macro-blaming language in the guide does IBM's damage at NOW's multiple, which is a much longer way to fall.

The Asymmetry Worth Respecting

NOW enters the print pre-damaged (the 7.5% sympathy hit lowered the bar) but still expensive (the premium multiple survived, so a genuine confirmation of the SaaS-erosion thesis has enormous room below). A clean print recovers the sympathy loss plus relief, call it high single digits. A dirty one re-rates the multiple itself, and multiple re-rates in software this year have run 20%+ (IBM just showed the template). The ±8% straddle splits the difference on a distribution that isn't symmetric.

The Options Angle

• If you own NOW, collar it through the print. Sell the call 8% up, own the put 10% down, roughly costless at current IV. You keep the recovery scenario's bulk and refuse the re-rate scenario entirely. This quarter, in this sector, unhedged long software through earnings is the one position with no defense.
• The clean speculative structure: put spreads, 10-20% down, expiring mid-August. The skewed tail is the mispriced side: sympathy selling already proved the market wants to believe the bear case, and confirmation would run further than relief. This is a hedge-sized bet on the trial going badly, not a core short.
• The acquittal trade waits for the print: if NOW posts receipts, buy the sector's relief through beaten-down peers (the sympathy victims: CRM, ADBE) with two-month calls, because they recover the contagion discount without having reported. Cheaper than chasing NOW itself up 8%.

Sooo... Does the Teacher's Pet Walk?

Lean no, with structure instead of size. The AI-eats-SaaS thesis has been right directionally all year, IBM gave it a body of evidence, and "premium execution" is a defense against bad quarters, not against a business-model question. NOW is the best-run company in the dock, which makes it the perfect test: if the charge sticks here, it sticks everywhere. Collar the longs, own the skew, and let July 22 read the verdict.