A Terrible, Terrible Week in Software Stocks. But Not Necessarily a Terrible 12 Months.

This week was bad. Really bad. And Thursday was particularly brutal.

ServiceNow was down 15-17% at the open. IBM off 10%. Salesforce down 7%. Oracle down 5%. Intuit down 7%. And the carnage spread to almost every major B2B name.

The trigger was earnings from two of the biggest legacy software names. Both beat. Both guided cautiously. And in April 2026, that’s all it takes to get obliterated.

What Actually Happened This Week

ServiceNow reported Q1 revenue of $3.77B, ahead of the $3.74B consensus. They still dropped 15%.

Why? A few reasons, none of them fatal:

• A 75 basis point subscription revenue headwind from on-premises deal slippage in the Middle East (~$23M), tied to the Iran conflict
• Full-year adjusted operating margin guidance cut from 32% to 31.5%
• Billings, a forward indicator, missed expectations
• No mention of their “generative AI book of business” metric, which sat at $12.5B at the end of 2025

IBM had its own problem. Revenue growth slowed to 9% from 12.2% the prior quarter. Software segment growth decelerated to 11.3% from 14%. That was enough for a 10% haircut, layered on top of the 13% single-day drop in February when Anthropic demonstrated Claude Code could modernize COBOL.

The read-through hit everything. Truist cut ServiceNow’s price target from $125 to $120 and wrote that “with heightened scrutiny on software vendors as frontier labs ramp enterprise revenue, the penalty for missteps becomes more severe.”

That’s the new reality. Beat and guide cautiously? Down 15%. Miss on anything? Down more.

The YTD Picture Is Genuinely Brutal

Pull up the year-to-date board and it’s a sea of red. Out of 25 major B2B names:

MNDY: -74.58%TEAM (Atlassian): -68.11%HUBS: -63.09%GTLB: -51.89%TTAN: -51.38%WDAY: -47.93%FIG (Figma): -46.76%NOW: -45.73%DOCU: -34.82%CRM: -34.56%SNOW: -34.08%RBRK: -33.00%SHOP: -22.99%TOST: -20.90%

Only DigitalOcean (+96%) is meaningfully green YTD. Almost every other name that touches B2B workflows is off 20% to 75% in under four months.

That’s what a sector-wide re-rating looks like. The iShares Expanded Tech-Software ETF (IGV) went from an all-time high of $117 to $85. Software collectively entered bear market territory in early February and hasn’t climbed out.

But Zoom Out to 12 Months and Something Different Shows Up

Switch the chart to 1-year and the picture changes materially. A lot of these same names are still up, in some cases massively:

DOCN: +233.88%NET (Cloudflare): +78.93%TWLO: +61.70%MDB (MongoDB): +55.77%PLTR (Palantir): +44.06%SHOP: +36.29%DDOG (Datadog): +32.48%CRWD (CrowdStrike): +15.57%

Even Snowflake, after all the carnage, is only -1.89% on a trailing 12-month basis. Palantir has been the single rare name that’s held up all year as its “Foundational Operating System” positioning gets re-rated as an AI enablement story rather than a disruption target.

The stocks that are down 12 months are mostly the ones the market has decided are most exposed to AI displacement risk: workflow tools, collaboration, document signing, anything where the core value prop is seat-based SaaS and the use case can be plausibly absorbed by an agent.

The stocks that are up are infrastructure, security, data, and AI-native platforms. Companies selling picks and shovels to AI, or companies whose product gets more valuable as agents proliferate.

What This Actually Means

The market isn’t saying software is dead. It’s saying two very specific things:

1. Seat-based pricing on human workflows is the exposed category. If your ARR scales with human logins and your product is a workflow an agent can replicate, your multiple is getting compressed. Period. This is why Monday, Atlassian, HubSpot, Workday, DocuSign, and ServiceNow are all getting hit.

2. Infrastructure beneath AI is the protected category. Cloudflare, MongoDB, Datadog, DigitalOcean, CrowdStrike are all up meaningfully because every new AI workload needs edge compute, databases, observability, and security. More agents means more infrastructure consumption, not less.

ServiceNow’s own response is instructive. They’ve introduced an “Agentic ACV” tier where customers pay for tasks completed by AI agents rather than per-seat. That’s the pivot every B2B company is being forced into right now, whether they want it or not.

AI Is Killing One Specific Business Model

A 12-month recovery in names like DOCN (+234%), NET (+79%), TWLO (+62%), MDB (+56%), and PLTR (+44%) means the narrative “AI is killing software” is too simple.

AI is killing one specific business model: charging $50-150 per seat per month for software that increasingly looks like something a model can do.

AI is minting a different business model: usage-based consumption of infrastructure, security, and data platforms that AI itself requires at 10x the prior scale.

The sorting is happening in real time. This week’s ServiceNow earnings reaction is part of that sorting. So was IBM’s COBOL moment in February. So was the broad Anthropic Cowork sell-off. The companies that figure out how to reprice, reposition, and turn agents from a threat into a wedge are going to look very different 18 months from now than they do in today’s red screen.

A Sector That Feels Dying, But Really is Just Being Rebuilt

If you’re holding B2B stocks right now, the YTD number is painful but not the most informative metric. The ones to watch are:

• Are they shifting from seat-based to outcome or consumption pricing?
• Is their “AI book of business” growing, shrinking, or quietly disappearing from the earnings deck?
• Is consumption up even if seat count is flat?
• Are they acquiring into AI-native categories or defending a shrinking moat?

Today was brutal. The 12-month chart tells you which names have already answered those questions and which ones are about to find out the hard way.

Either way, this isn’t a sector dying. It’s a sector being rebuilt with very different rules, at very different multiples, with a very different set of winners.