AI Models & Agents

IBM Stock Plummets 13% After Anthropic Targets COBOL AI

Anthropic’s new Claude AI tool threatens IBM’s legacy mainframe business, triggering its biggest drop since 2000.

By Kronos News Desk··1 min read
A photograph of a modern data center with glowing blue server racks. A transparent digital graphic of a declining stock market line and snippets of computer code are superimposed over the image.

A photograph of a modern data center with glowing blue server racks. A transparent digital graphic of a declining stock market line and snippets of computer code are superimposed over the image.

Photo: Kronos News

Shares of International Business Machines (IBM) plunged more than 13% on February 23, 2026, wiping out approximately $31 billion in market value [2][4]. The sell-off followed an announcement from AI startup Anthropic regarding new capabilities for its Claude Code tool to automate the modernization of COBOL systems [1][3]. Investors reacted to the potential disruption of IBM’s lucrative mainframe consulting business, which manages legacy infrastructure for 95% of U.S. ATM transactions [2][4]. ## The Challenge to Legacy Computing Anthropic claims its AI can analyze and document complex COBOL codebases in quarters rather than the years typically required by human consultants [1]. By identifying system dependencies and mapping workflows, the tool aims to lower the barrier for organizations to migrate from decades-old mainframes to modern cloud environments [1][3]. While IBM has its own AI conversion tools, the market viewed Anthropic’s update as a direct threat to the specialized consulting model that sustains IBM’s legacy profitability [3]. IBM responded by emphasizing that while AI can simplify code translation, the true difficulty lies in data architecture and transaction integrity [1]. A spokesperson noted that IBM’s deep hardware-software coupling remains a significant hurdle for any automated migration tool [1]. Despite this defense, the stock’s decline marked its steepest single-day loss since October 2000 [2][4].

Editorial notes

Transparency note

Drafted with LLM; human-edited

AI assisted
Yes
Human review
Yes
Last updated

Risk assessment

Under review

Reviewed for sourcing quality and editorial consistency.

Sources

Related stories

View all

Topics

Get the weekly briefing

A concise briefing with selected stories and analysis.

No spam. Unsubscribe anytime. By joining, you agree to our Privacy Policy.

About the author

Kronos News Desk covers ai models & agents and editorial analysis for Kronos News.