Impact of Claude Code on IBM's COBOL Business

Claude, Anthropic's AI solution, has significantly impacted the perceived value of IBM's COBOL and mainframe business in the short term, primarily by emphasizing the risks associated with AI-driven modernization. While immediate market reactions have been largely negative, the long-term implications for COBOL-related revenue growth are more complex and debated among analysts.

COBOL remains essential for legacy systems in various sectors, including finance, government, and airlines. It processes a substantial volume of transactions—approximately 95% of U.S. ATM transactions—on IBM's mainframes, known as IBM Z. For years, IBM has profited from high-margin hardware, software, services, and consulting that support the maintenance and modernization of these extensive, decades-old codebases. However, modernization presents challenges, such as dealing with "spaghetti code," numerous dependencies, a shortage of experienced professionals, and the critical need for reliability and security.

On February 23, 2026, Anthropic released a blog post and a "Code Modernization Playbook" that showcased Claude Code's capabilities concerning COBOL. The tool automates exploration and analysis of code, mapping dependencies, tracing execution paths, and documenting workflows. It also facilitates incremental refactoring and migration to modern languages and cloud environments, claiming to shorten multi-year projects to just a few quarters by minimizing the costly upfront analysis phase.

Following this announcement, IBM's stock plummeted approximately 13% in a single day—the largest drop since October 2000—resulting in a loss of around $30 billion in market capitalization. The overall decline in February was about 27%, with related companies like Accenture and Cognizant also experiencing losses. Investors were concerned about the potential erosion of IBM's high-margin infrastructure services and consulting business, which has historically thrived due to the mainframe "flywheel" effect.

In response, IBM quickly countered the narrative through a blog by Rob Thomas. The company argued that modernization involves more than just code translation; it encompasses platform complexities, performance, security, and integration—all areas where mainframes excel. IBM maintains that AI tools, including their Watsonx and Project Bob, enhance the case for IBM Z by streamlining refactoring, DevOps, and knowledge preservation while addressing skills gaps.

Analysts emphasize that while Claude assists in analysis and refactoring, true modernization requires addressing more than syntax, including business logic, testing, compliance, and data flows. Many enterprises prefer to remain on proven mainframe systems, and IBM's own AI offerings target the same modernization use cases.

Overall, the short-term outlook for COBOL's growth at IBM appears negative, with market sentiment highlighting disruption risks. However, in the medium to long term, the impact remains uncertain. AI may lower barriers to modernization, potentially accelerating off-mainframe migrations, but it also presents opportunities for IBM to market more AI-enhanced mainframe services. While COBOL may not be experiencing growth in new development, its established base remains crucial, and AI can help manage it effectively without rendering it obsolete.