Skip to main content Scroll Top

IBM faces a major crisis as 95 percent of COBOL ATMs at risk from AI disruption

WHY THIS MATTERS IN BRIEF

AI that modernises COBOL wiped 13% off IBM's stock in a single day.

 

Matthew Griffin is the World’s #1 Futurist Keynote Speaker and Global Advisor for the G7 and Fortune 500, specialising in exponential disruption across 100 countries. Book a Keynote or Advisory SessionJoin 1M+ followers on YouTube and explore his 15-book Codex of the Future series.

 


 

First it was SaaS stocks, then cybersecurity firms’ shares, and now “Big Blue” is in the crosshairs: IBM’s stock recorded its strongest single-day loss in over 25 years recently, and the trigger was the announcement by AI startup Anthropic that its tool Claude Code could significantly simplify the modernisation of COBOL systems. The stock plummeted 13.2 percent, the highest decline since October 2000. Investors fear that Artificial Intelligence (AI) could fundamentally change IBM’s traditional mainframe business.

 

RELATED
Nanomagnetic computing could help create ultra efficient AI's at the edge

 

As reported, the market recently observed that SaaS stocks came under pressure following the release of Claude Cowork, and then Claude Code Security caused a shock in cybersecurity stocks.

COBOL (Common Business-Oriented Language) is a programming language developed in the late 1950s. Despite its age, it still runs on countless systems worldwide today. By some estimates, 95 percent of all ATM transactions in the United States are processed via COBOL. Hundreds of billions of lines of COBOL code are in production daily, controlling mission-critical systems in banks, insurance companies, airlines, and government agencies.

The language runs primarily on mainframes, many of which are from IBM. These powerful server systems are particularly in demand in regulated industries where reliability and stability are top priorities. The problem: the developers who built these systems are long retired, and COBOL is taught at only a few universities anymore. The number of specialists who understand and can maintain the code is shrinking continuously.

 

RELATED
OpenAI has a new plan to replace UBI with a Soverign AI Wealth Fund

 

 

The Future of Vibe Coding and SAAS Keynote by Futurist Keynote Speaker Matthew Griffin

Book a Keynote with Matthew

 

In a blog post, Anthropic explained that “Claude Code” could help companies modernise COBOL systems. The tool could automate tasks that previously accounted for the bulk of the effort such as mapping dependencies across thousands of lines of code, documenting workflows that no one remembers anymore, identifying risks that would take human analysts months to uncover and providing detailed insights for informed decision-making

“Modernizing a COBOL system used to require armies of consultants mapping workflows for years. With AI, teams can modernise their COBOL codebase in quarters instead of years.”

This statement hit IBM at a sensitive spot. The mainframe business is a key revenue pillar for the company. If AI tools simplify the migration of COBOL systems, this could impact demand for IBM mainframes in the long term.

IBM’s stock crash is part of a larger trend. Software stocks have been under pressure since the start of the year. A major software ETF (IGV) is down 27 percent and is heading for its strongest quarterly loss since the 2008 financial crisis.

 

RELATED
Quadruple stranded DNA seen in human cells for the first time

 

Cybersecurity companies like CrowdStrike (down 9.85 percent) and Datadog (down 11.28 percent) also came under selling pressure on Monday after Anthropic unveiled a new security feature for its Claude AI model.

Investors are increasingly concerned about “vibe coding,” the ability to generate software code independently using AI. The fear: if users can increasingly create their own applications, demand for established software products will decline.

The idea of making COBOL more accessible using modern AI tools is not new. IBM itself launched an AI-powered tool in 2023 that analyses COBOL code and can convert it to programming languages like Java, among other things.

IBM CEO Arvind Krishna emphasised in July 2025 that the AI coding assistant for mainframes had “achieved very broad adoption.” In most cases, customers used the tool to better understand their COBOL codebase and decide what should be modernised.

 

RELATED
AI developer Anthropic hires a Welfare Officer ... for its AI

 

Whether the market reaction is justified remains to be seen. IBM has lost a total of 27 percent in value in February and is heading for its strongest monthly decline since at least 1968. While investors react to disruption risks in the short term, IBM is strategically betting on AI integration in the mainframe environment itself.

The central question remains: will customers actually migrate their proven COBOL systems just because it becomes technically easier?

 


 

Why would AI modernising COBOL hurt IBM so much?
COBOL still runs on IBM mainframes across banks, insurers and government, and modernising it has long needed armies of consultants; if AI tools let teams migrate that code in quarters instead of years, demand for IBM's mainframes – a key revenue pillar – could shrink over time.

Related Posts

Leave a comment

Pin It on Pinterest

Share This