Three unrelated events collided in a single trading session — and sent IBM stock into freefall, wiping out a quarter century of gains in one day.
| -13.2% IBM Single-Day Drop | -27% IBM Drop in February | $223.35 Closing Price (Feb 23) |
On Monday, February 23, 2026, IBM suffered its single worst trading day since the dot-com crash of October 2000. Shares plunged 13.2%, closing at $223.35, and capping off a brutal February that has now erased 27% of the company’s market value — its worst monthly performance since at least 1968. But this was not a story of bad earnings or a corporate scandal. It was the story of three separate forces — a viral research report, an AI product announcement, and a market strategist’s warning — that crashed into each other at precisely the wrong moment.
Trigger #1: The Citrini Research “2028 AI Crisis” Report
It started over the weekend with a social media post. Citrini Research, a relatively unknown firm founded by James van Geelen, published a report titled “The 2028 Global Intelligence Crisis” — a hypothetical scenario analysis exploring what a world of aggressive AI adoption might look like just two years from now. The original tweet garnered 4.5 million views, 2,100 retweets, and 12,000 bookmarks within hours.
The report laid out a dystopian-but-plausible scenario set in June 2028 where AI disruption has triggered mass white-collar unemployment, declining consumer spending, software-backed loan defaults, and broad economic contraction. The authors were explicit: “What follows is a scenario, not a prediction.” But markets didn’t care about the disclaimer.
“We are certain some of these scenarios won’t materialize. As investors, we still have time to assess how much of our portfolios are built upon assumptions that won’t survive the decade.”
— Citrini Research Report, Feb 23, 2026
Among the disruption scenarios Citrini outlined: food delivery services like DoorDash and Uber Eats displaced by AI-coded alternatives; payment processors like Mastercard and Visa seeing their transaction fees eliminated by AI agents optimizing for cost savings; and private credit firms facing defaults as AI erodes the revenue bases of their portfolio companies.
Trigger #2: Anthropic’s Claude Code Targets IBM’s Crown Jewel
While investors were still absorbing the Citrini report on Monday morning, Anthropic dropped a blog post with a very specific and very concrete threat to IBM’s core business: Claude Code can now modernize COBOL.
What is COBOL and why does it matter to IBM?
COBOL (Common Business-Oriented Language) is a programming language developed in 1959. It sounds ancient — because it is. Yet today, COBOL powers an estimated 95% of ATM transactions in the United States, along with critical systems in banking, airlines, and government. The reason these systems haven’t been replaced is simple: modernizing COBOL has historically been too expensive and too risky. The code is notoriously complex, the pool of developers who understand it is shrinking every year, and the systems are too mission-critical to break.
IBM has built an enormous and highly profitable business on this reality. Its mainframe hardware, the Z-series, runs the majority of the world’s COBOL systems. IBM generates consulting revenue, software revenue, and hardware revenue from maintaining, modernizing (on its own terms), and keeping these clients locked into its ecosystem. IBM’s CFO James Kavanaugh described the flywheel plainly after the company’s last earnings: the Z mainframe generates a 3x-4x revenue stack multiplier across IBM’s entire business.
What Anthropic’s Claude Code Does
Anthropic’s blog post claimed that Claude Code can automate the most time-consuming and costly phases of COBOL modernization — the analysis and exploration phases that have historically required expensive human consultants spending months mapping out code dependencies. Specifically, Claude Code can map dependencies across thousands of lines of code, document hidden workflows, identify risks, and provide engineers with the insight needed to redesign systems for modern cloud infrastructure. Anthropic’s claim: teams can now modernize COBOL codebases in quarters, not years. The company also released a Code Modernization Playbook alongside the announcement.
| Collateral Damage: Who Else Got Hit IBM: -13.2% (worst day since Oct 2000)Accenture & Cognizant: Both declined sharply (major COBOL consulting practices)DoorDash & American Express: Down 6.5%+Mastercard, Visa, Capital One, Apollo, Blackstone: Down 4-6%Software-focused ETFs: Down 4.8%S&P 500: Down 1% |
Trigger #3: Nassim Taleb Adds Fuel to the Fire
As if Citrini and Anthropic weren’t enough, renowned author and risk analyst Nassim Nicholas Taleb issued a public warning on Monday: investors should brace for escalating volatility and even bankruptcies in the software sector as the AI rally enters what he called a fragile phase. Taleb’s framing — that markets are underpricing structural risk while overestimating the durability of current AI leaders — added intellectual credibility to an already-panicked narrative.
Is IBM Actually Dead? Not So Fast.
The market reaction may have been dramatically outsized relative to the actual threat. IBM is not standing still — it has its own AI tool for COBOL modernization (watsonx), and crucially, it wants the outputs to stay running on its mainframe hardware. Clients with decades of COBOL infrastructure may prefer the security and reliability of IBM’s known ecosystem over migrating to untested cloud infrastructure.
Thomas George, a portfolio manager at Grizzle Investment Management, put it well: “The report raises real concerns about disruption, even if things don’t end up as dire as the worst-case scenario. Certainly you don’t feel great after reading it, and I’m sure it leaves anyone holding these stocks with less conviction.”
The broader context matters too. IBM has been under pressure throughout February amid wider anxieties about AI’s disruption of enterprise tech. Citrini’s report and Anthropic’s announcement didn’t create the fear — they amplified an existing, raw nerve in the market.
The Bigger Picture: The “AI Scare Trade” Is Real
What happened to IBM on Monday is part of a wider market phenomenon that analysts are calling the “AI Scare Trade.” In recent weeks, sectors as varied as software, cybersecurity, wealth management, insurance brokerage, logistics, and private credit have all been caught in indiscriminate selloffs triggered by AI announcements. The pattern is consistent: an AI company announces a new capability, investors immediately identify which incumbents are threatened, and the stocks drop before any evidence of actual disruption materializes.
Just last Friday, a new Claude Code Security feature caused a selloff in cybersecurity stocks. Before that, other AI product announcements triggered drops in different sectors. The market is effectively repricing legacy tech in real time — and doing it fast, often before the dust settles.
