IBM shares fell yesterday more than 13% due to investor reaction to an Anthropic blog post claiming that new tools from its AI assistant Claude Code can be used to modernize legacy systems running COBOL.
IBM is the latest collateral victim of the possibilities of new artificial intelligence functions. As happened in the field of security software with the announcement of Claude Code Security, the simple publication of the article “How AI helps break the cost barrier to modernize COBOL,” has caused an escalation in the sale of securities that also affected other companies such as Accenture y Cognizant.
They all have legacy systems modernization solutions that generate revenue by helping organizations upgrade decades-old COBOL systems. And they are all being penalized in the face of the capabilities of what is coming from artificial intelligence, in changes in business software that are presumed historic, whether in cybersecurity or, as is the case at hand in programming.
The decline in IBM’s share Price comes at a time when there is speculation that AI will ruin SaaS-based business modelsan idea that is believed to be behind substantial decreases in the capitalization of other large software companies: Salesforce, Atlassian, Adobe, ServiceNow or HubSpot, among others.
IBM and COBOL actions
COBOL, short for Business Oriented Common Languageis a programming language developed in the late 1950s. Despite its longevity, Its importance is vital in global technology. Hundreds of billions of COBOL lines run in production every day, powering critical systems in the financial, airline, or government sectors. Just one piece of information to understand its importance: it is estimated that the language manages 95% of the transactions made at ATMs in the United States.
IBM has been selling systems for a long time mainframe optimized for large-scale transaction processing, especially under COBOL. The blue giant assures that these great machines “still offer the lowest operating costs for some workloads” and it is worth remembering that the same company use AI code conversion tools from 2023.
However, COBOL continues to be a burden for many organizations due to the high bills they must pay to keep such an old language working. Anthropic wants to be part of the solution. The AI firm notes that it is difficult to find COBOL-proficient programmers, that attempts to train them have failed to increase their numbers, and that migrating from COBOL is therefore risky and expensive. Company suggests using its Claude Code AI tool to help rewrite COBOL applications
“AI can evaluate which components are safe to move and which require careful handling”they explain. “Areas with accumulated technical debt are documented before they become migration surprises”. At first glance, the publication portends trouble for IBM’s mainframe business and hence the stock’s decline.
The opportunities presented by migrations to COBOL and potential of AI to accelerate refactoring of legacy applicationsare not new to the technical community. The rise of AI means that the cost of rewriting legacy applications has become affordable, making these measures imperative.
Claude Code can help modernize COBOL codebases by mapping dependencies across thousands of lines of code, documenting workflows, and identifying risks that “it would take human analysts months to discover”they say Anthropic. If modernizing legacy code stalled for years because understanding it cost more than rewriting it, «AI reverses that equation»they assure.
