IBM shares recorded their steepest one-day fall in more than 25 years after fresh concerns emerged about how artificial intelligence could disrupt one of the company’s most dependable revenue streams.
The stock plunged 13.2 percent to close at $223.35, marking its worst single-day decline since October 18, 2000. So far this year, IBM shares have fallen about 25 per cent as investors reassess how quickly AI could reshape the economics of enterprise software and IT services.
The latest sell-off followed a blog post from AI startup Anthropic. The company claimed that its AI tool, called Claude Code, can understand and modernise COBOL, a programming language developed in the 1950s that still powers many of the world’s critical systems.
COBOL remains deeply embedded in banks, airlines, insurance firms, and government agencies. It also plays a central role in IBM’s mainframe business. For decades, updating COBOL systems has been slow, costly and heavily reliant on teams of specialised consultants. This steady demand has provided IBM with reliable revenue, as many organisations struggle to maintain legacy systems that few engineers fully understand today.
Anthropic argues that AI could dramatically change that dynamic. According to the company, there are hundreds of billions of lines of COBOL code still running in live systems every day, even as the number of programmers skilled in the language continues to shrink.
The startup says AI is particularly effective at handling the complex and time-consuming tasks that once made modernising COBOL systems too expensive. It estimates that around 95 per cent of ATM transactions in the United States still rely on COBOL, highlighting how deeply the language is woven into financial infrastructure.
Anthropic claims its AI can scan massive codebases, map out how different parts of a system connect, generate clear documentation for poorly understood software, and flag potential risks that would otherwise take months to identify manually.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
The company summed up its position clearly, saying that modernisation efforts have stalled for years because understanding old code often costs more than rewriting it. With AI, it argues, that equation is finally starting to change.
