Why did IBM shares plunge after Anthropic's COBOL claim?
Market reaction to a new AI tool that targets legacy systems
Investors sold IBM stock sharply after Anthropic released a public demonstration showing that its Claude Code tool can automate large parts of the exploration and analysis phases of COBOL modernization. Traders interpreted that demonstration as evidence that AI is beginning to commoditize a set of services — assessment, code analysis, and initial modernization planning — that have long been profitable for incumbents like IBM.
The sell-off reflects several linked concerns. First, IBM earns substantial revenue from consulting and software that helps governments and enterprises migrate decades-old mainframe applications. If advanced AI can quickly analyze COBOL code, map dependencies, and propose translations, that work becomes less labor-intensive and—critically—more contestable by newer entrants offering automated tools.
Second, the timing and public nature of Anthropic’s post accelerated market anxiety. Even though IBM has for years pointed to automation and tooling for legacy code modernization, investors reacted to the possibility that Anthropic’s offering could undercut pricing or reduce the scale of human-led engagements.
Key points
- The drop reflected fears that a core revenue stream could be disrupted by AI-driven automation.
- The market response highlights how quickly investor sentiment can shift when AI tools demonstrate domain-specific capabilities.
- Uncertainty remains about accuracy, auditability, and the downstream costs of using AI to modernize mission-critical systems.
Analysts caution that demonstrated capability is not the same as enterprise-grade product adoption. Large-scale modernization projects still demand thorough testing, verification, and long-term support — areas where established integrators retain advantages. Nonetheless, the episode marked a clear signal that AI developments are reshaping investor expectations about legacy IT services.