IBM Shares Tumble After Anthropic Claims AI Breakthrough in COBOL Modernisation

Shares of IBM recorded their steepest single-day decline in more than 25 years after AI startup Anthropic announced that its Claude Code tool can understand and modernise COBOL — a programming language central to IBM’s mainframe ecosystem.

IBM’s stock closed 13.2% lower at $223.35, marking its sharpest drop since October 18, 2000. The fall has extended the company’s losses to roughly 25% so far this year, reflecting growing investor unease about how rapidly artificial intelligence could reshape long-standing enterprise technology models.

The trigger was a blog post from Anthropic highlighting the capabilities of its AI-powered coding assistant, Claude Code. The company said the tool can interpret and streamline massive COBOL codebases — a task that traditionally required large consulting teams and multi-year modernisation projects, areas where IBM has historically generated steady revenue.

COBOL (Common Business-Oriented Language), developed in the late 1950s, remains deeply embedded in banking systems, insurance platforms and government infrastructure worldwide. IBM’s mainframes, designed for high-volume transaction processing, continue to rely heavily on the language. Anthropic estimates that about 95% of ATM transactions in the United States still depend on COBOL, underlining its enduring importance.

In its blog, Anthropic emphasised the scale of the challenge. “Hundreds of billions of lines of COBOL run in production every day, powering critical systems in finance, airlines, and government. Despite that, the number of people who understand it shrinks every year,” the company wrote. “AI excels at streamlining the tasks that once made COBOL modernisation cost-prohibitive.”

The startup said Claude Code can trace dependencies across thousands of lines, generate documentation for legacy workflows and flag potential risks that might otherwise take months of manual review. “Legacy code modernisation stalled for years because understanding legacy code costs more than rewriting it. AI flips that equation,” Anthropic said.

It further claimed that projects once measured in years could now be completed much faster. “With AI, teams can modernise their COBOL codebase in quarters instead of years,” Anthropic said. That assertion appears to have rattled investors concerned that automation could compress timelines and reduce demand for consultant-heavy transformation programmes.

The reaction reflects broader anxiety across global IT markets. Technology stocks have faced sustained pressure as AI tools move quickly from pilot phases into real-world deployment. Anthropic’s push to position Claude as an application layer capable of automating complex software tasks has intensified concerns that traditional consulting and integration services could see pricing pressure.

The unease extends beyond the United States. Indian IT companies have also experienced volatility amid debate over whether AI-driven automation will reduce workforce-heavy delivery models.

Not everyone sees disruption as a threat. At Wipro, Chief Strategist and Technology Officer Hari Shetty suggested AI could expand opportunities rather than diminish them. “When you look at the entire gamut of things that's possible, it really appears like a large opportunity for us,” he said.

However, former Infosys CEO Vishal Sikka has cautioned that the shift is already underway. “If you look at the application of generative AI to knowledge work, this disruption is real. It is here,” he said, pointing to productivity gains in code migration and system integration.

For investors, the message is clear: AI is no longer theoretical. It is beginning to challenge even the most entrenched corners of enterprise computing.