FEATURED

The SaaSpocalypse Made a $2 Trillion Mistake

Here is the uncomfortable math. ServiceNow is growing subscription revenue at 22% year over year. Salesforce just crossed $1.2 billion in Agentforce annual recurring revenue, up more than 200% in a year. Both companies are generating billions in free cash flow. And both stocks have been absolutely punished in 2026.

ServiceNow is down roughly 33-34% for the year. Salesforce has dropped approximately 38-41% and hit a fresh 52-week low in late June. The broader software index has lost around 20% in the same window. Wall Street has a name for it: the SaaSpocalypse.

The logic goes like this: AI agents automate workflows. Enterprise software companies charge per seat for people running those workflows. Fewer human seats means less revenue. Sell everything with a SaaS business model before the machines take it all.

The fear is not irrational. What it may have gotten wrong is which companies it actually applies to.

What the Market Actually Priced

The SaaSpocalypse accelerated sharply in early 2026 after AI agent products from companies like Anthropic and OpenAI started being positioned as direct enterprise software competitors. The selloff that followed wiped hundreds of billions from software market caps over a matter of weeks. For the first time in recent memory, software forward price-to-earnings multiples fell below the broader S&P 500 average. The market applied a blanket discount to every SaaS name regardless of what it actually does.

That is the category error.

There is a meaningful difference between a company that sells software for tasks AI agents could plausibly replace, and a company whose entire function is to govern, audit, and orchestrate what those agents do inside large organizations. ServiceNow falls into the second category. The platform processes more than 100 billion workflows annually and has accumulated two decades of enterprise context. Every AI agent a company deploys needs to be governed, tracked, and audited across systems. That is the infrastructure ServiceNow provides.

Salesforce sits in a different but equally defensible position. Its Data Cloud unifies data across siloed systems, and AI agents are only as useful as the data they can access. The Agentforce platform is built on top of that data layer, giving enterprises a unified record from which to actually run AI workflows. You cannot easily replicate that institutional data context by building a new tool from scratch.

The Numbers the Market Is Not Reading

ServiceNow’s Q1 2026 results beat the high end of guidance across every major metric. Subscription revenues reached $3.671 billion, representing 22% year-over-year growth (19% in constant currency). Current remaining performance obligations stood at $12.64 billion, up 22.5% year over year. Now Assist customers spending over $1 million in annual contract value grew more than 130% year over year. The company also raised its full-year AI revenue target from $1 billion to $1.5 billion, with Now Assist ACV tracking at roughly $750 million as of Q1.

The company also reported 16 transactions over $5 million in net new ACV in Q1 alone, representing nearly 80% year-over-year growth. New logo ACV growth accelerated to over 50% year over year. That is not a business model under existential threat. That is a business compounding at an unusual rate while its stock has been cut by a third.

Salesforce is a different conversation but a similar dislocation. Q1 fiscal 2027 revenue came in at $11.13 billion, ahead of the $11.05 billion consensus. Non-GAAP EPS of $3.88 beat estimates by approximately 24% and was up 50% year over year. Agentforce ARR reached $1.2 billion, up 205% year over year, with combined Agentforce and Data 360 ARR hitting roughly $3.4 billion. The company’s current remaining performance obligation rose 14% to $33.6 billion. The board had authorized a $50 billion share repurchase program, and in Q1 the company executed a $25 billion accelerated share repurchase, driving the diluted share count down 10% year over year.

And yet. Salesforce trades near multi-year lows. ServiceNow’s valuation has compressed to levels not seen during its entire run as a public company. The multiple destruction has been severe, and it has been indiscriminate.

What Institutions Are Starting to See

On July 1, Guggenheim upgraded both Salesforce and ServiceNow to Buy, citing the disconnect between current valuations and underlying business performance. Analyst John DiFucci argued that current prices imply a scenario where Salesforce “declines 5% into perpetuity” — a view he called misaligned with reality. His ServiceNow price target of $125 and Salesforce target of $228 both reflect pure valuation recovery, not a bet on AI dominance. The upgrades sent both stocks up roughly 4-5% on the day.

Slight tangent, but it matters: the SaaSpocalypse fear assumes AI replaces enterprise workflows cleanly and quickly. What every CIO knows is that enterprise AI deployment is slow, messy, and governed by compliance requirements that do not disappear because a model gets better at reasoning. The companies that own the compliance and governance layer of AI deployment are structurally positioned to benefit from adoption acceleration, not be destroyed by it. That case does not seem to be getting through to the market yet.

What is happening in early July is a quiet but accelerating reassessment of the entire enterprise software complex. It is a fundamental recalibration — institutional capital beginning to acknowledge that the market mispriced disruption risk while ignoring durable revenue growth.

What Comes Next

ServiceNow reports Q2 2026 results on July 22. Management guided subscription revenue of $3.815 to $3.820 billion for the quarter, implying roughly 21-21.5% constant currency growth. The question the market will be watching is whether Now Assist ACV continues its trajectory toward the $1.5 billion full-year target, and whether the Armis acquisition integration stabilizes margins or creates further headwinds.

Three simultaneous acquisitions — Moveworks, Armis, and Veza — is a genuine execution risk. Large enterprise software M&A historically consumes management attention in ways that slow organic growth. That is the real risk here, not AI displacement. The stock fell 17% after Q1 results despite beating guidance, largely because margin guidance disappointed after the Armis deal closed. A clean Q2 could change that reaction function entirely.

For Salesforce, the next catalyst is fiscal Q2 FY27 earnings, where investors will watch Agentforce paid seat additions at scale for the first time. The Fin acquisition for $3.6 billion, announced June 15, adds a 30,000-company customer base and a proprietary AI model purpose-built for customer support, extending the data advantage that underpins the entire Agentforce story.

Options Structure Worth Watching

For traders expecting a re-rating in enterprise software ahead of ServiceNow’s July 22 report, implied volatility is elevated following the extended selloff. Defined-risk structures that benefit from a move higher without requiring precise timing include bull call spreads in the August expiration cycle. For traders skeptical that one strong quarter resolves the structural multiple compression, put spreads below current support remain a clean expression of continued uncertainty.

The bull case: growth holds above 20%, Now Assist ACV trajectory confirms AI monetization is real, and the multiple re-rates toward historical mid-range, implying significant upside from current levels. The bear case: Armis integration disrupts margin guidance, a macro deceleration in enterprise IT spend slows large deal closings, and the stock revisits April lows near $81. The neutral case: earnings meet guidance, the stock stabilizes but does not accelerate, and the re-rating waits for a second consecutive clean quarter.

The market made a category error in early 2026. The question now is not whether that error happened — the valuation data confirms it did. The question is when institutional capital fully corrects it. Q2 reporting season, kicking off July 22 with ServiceNow, is the most likely forcing function.