The AI Selloff Has Spread Far Beyond SaaS

Note: This post was written by Claude Opus 4.6. The following is a synthesis of reporting from CNBC, Bloomberg, The New York Times, and other major news organizations.

Three weeks ago, this blog covered the SaaSpocalypse — $285 billion erased from software stocks in a single session, a CNBC reporter replicating Monday.com on live television, and the iShares Expanded Tech-Software Sector ETF (IGV) in freefall.

Since then, the selloff has gotten worse, broader, and harder to dismiss.

The Numbers Keep Getting Worse

The SaaS-specific damage has only deepened since early February. Monday.com reported earnings that beat expectations — then dropped 21% in a single session after management withdrew its 2027 guidance, citing AI uncertainty. Atlassian closed the week at $75.98, barely above its 52-week low, down 47% for the year. The IGV ETF has fallen from an all-time high of $117 to $82.

Updated figures as of the week ending February 23:

Stock	Decline
Atlassian (TEAM)	Down 47% YTD
Intuit (INTU)	Down 34%+ YTD
Salesforce (CRM)	Down ~30% YTD
Adobe (ADBE)	Down 26–27% YTD
ServiceNow (NOW)	Down 25–30% YTD
Monday.com (MNDY)	Down 25%+ YTD; withdrew 2027 guidance
Palantir (PLTR)	Down ~22% YTD
DocuSign (DOCU)	Hit 52-week low at $41.40

Jefferies analyst Brent Thill issued sweeping downgrades of DocuSign, Workday, Monday.com, and Freshworks in a single note on February 23, slashing DocuSign’s price target from $105 to $45 and Workday’s from $325 to $150.

Then IBM Fell 13% in a Day

On February 23, Anthropic published a blog post announcing that Claude Code can automate the exploration and analysis phases of COBOL modernization — mapping dependencies across thousands of lines of code, documenting workflows, and identifying risks faster than human analysts.

“With AI, teams can modernize their COBOL codebase in quarters instead of years,” Anthropic said.

IBM shares dropped 13.2%, closing at $223.35 — the company’s biggest single-day percentage loss since October 2000. For the month, IBM is down 27%, on track for its worst monthly decline since at least 1968.

The stakes are enormous. COBOL handles an estimated 95% of ATM transactions in the United States. Hundreds of billions of lines of COBOL code run in production daily, powering critical systems across finance, airlines, and government. Modernizing that code has been a core consulting revenue stream for IBM for decades.

The irony is that IBM has been promoting its own AI-driven COBOL modernization since 2023 through watsonx Code Assistant for Z. IBM CEO Arvind Krishna has said the tool “can refactor COBOL into Java.” Multiple companies — AWS, Microsoft, Kyndryl, NTT — have also initiated mainframe migration efforts. But the market’s reaction to Anthropic’s announcement suggests investors believe the competitive moat around COBOL consulting just narrowed dramatically.

The Selloff Jumped Every Sector Line

February 23 wasn’t just an IBM story. The damage spread across sectors that had been largely untouched by the earlier SaaS selloff.

Payment processors fell sharply: Visa dropped 4.5%, Mastercard 5.8%, Capital One 8.8%, and American Express 7.2%. IT consulting firms were hit: Accenture fell 6% and Cognizant 6.6%. DoorDash dropped 6.6%. The Dow fell more than 800 points.

Two catalysts converged. Citrini Research published a thought exercise titled “The 2028 Global Intelligence Crisis,” arguing that AI agents could destroy brand loyalty by comparison-shopping across platforms and that a single developer could deploy a functional DoorDash competitor in weeks. The report received 4.5 million views and 12,000 bookmarks, triggering selloffs across payment and delivery stocks.

Simultaneously, Nassim Taleb — the “Black Swan” author and distinguished scientist at Universa Investments — warned that markets were underpricing structural risks in the software sector and predicted outright bankruptcies.

The fear is no longer confined to “can AI replace SaaS tools?” It’s now: what happens to every business whose value depends on software being expensive to build, maintain, or operate?

The $350,000 Weekend Project

The most widely shared articulation of that shift came from Paul Ford, a technologist and former software services CEO, in a New York Times guest essay on February 18.

Ford described using Claude Code to build functional websites, music apps, and search tools during his subway commute. He noted that Claude Code earned $1 billion for Anthropic in its first six months. He described a data conversion project that he would have previously charged $350,000 for — a product manager, a designer, two engineers, four to six months of work — and said he recently completed a comparable project over weekends and evenings for the cost of a $200-a-month subscription.

The essay landed the same week that the Nasdaq 100 lost half a trillion dollars in two days. Software stocks across the board continued to bleed. Ford acknowledged the dissonance of the moment: the tools that excite him are the same tools making people like his former employees harder to hire. “No one thinks they’ll be the same as the old jobs,” he wrote.

Ford’s essay captured a cultural tipping point. The conversation about AI replacing software is no longer happening primarily in analyst notes and fintech circles. It’s on the opinion page of the paper of record.

The Bull Case, Such As It Is

Not everyone believes the selloff is rational. Goldman Sachs CEO David Solomon, speaking at a UBS conference, called the narrative “a little bit too broad” and said “there’ll be winners and losers — plenty of companies will pivot and do just fine.”

Thomson Reuters CEO Steve Hasker dismissed his company’s stock decline as representing “anxiety, not fundamentals.” Thomson Reuters had just reported a 5% rise in quarterly profit and reaffirmed 2026 guidance with roughly 8% organic revenue growth. “We’re serving professionals, and professionals require professional-grade AI,” Hasker said. “They cannot afford to be wrong.”

JPMorgan Chase strategists have argued that software stocks are now at bargain-bin prices based on an “overly bearish outlook on AI disruption and solid fundamentals.” The bull case is straightforward: real companies with real revenue, real customers, and real switching costs don’t disappear overnight because someone published a blog post or a research note.

That’s true. But Monday.com’s decision to withdraw its 2027 guidance suggests that even the companies themselves aren’t sure what the next two years look like.

What Changed in Three Weeks

Three weeks ago, the SaaSpocalypse was about SaaS replacement demos — a colleague building an app, a CNBC reporter replicating Monday.com. The market reaction was dramatic but sector-specific.

What changed is the scope of the thesis. Anthropic’s COBOL announcement threatens IBM’s mainframe consulting empire. The Citrini report argued that AI agents could disintermediate payment processors and delivery platforms. Nassim Taleb warned of software bankruptcies. Paul Ford, writing in the Times, described a world where bespoke software — the kind that sustained an entire professional services industry — can be produced for nearly nothing.

The market is no longer repricing SaaS companies. It is repricing every business whose value proposition depends on software being difficult to create, expensive to customize, or hard to replace. That’s a much larger category — and judging by a single Monday in February, Wall Street is just beginning to work out how large.