IBM experienced a significant decline in its share price, dropping 13 percent on Monday, largely attributed to investor reactions to a blog post by Anthropic. The AI company highlighted its Claude Code tools, which it claims can enhance the refactoring process of applications written in COBOL, a programming language known for its longevity in the tech world.
In the blog post, Anthropic emphasized that COBOL applications are still widely used, playing critical roles in sectors such as government, finance, and airlines. However, the blog also pointed out a major challenge: the shortage of programmers skilled in COBOL. Efforts to increase the number of professionals familiar with the language have not yielded substantial results, making the transition away from COBOL both risky and costly.
Anthropic proposed that their AI tools could assist in the migration of COBOL applications by identifying which components of the software are suitable for migration and which require more careful attention. They highlighted the importance of documenting areas with accumulated technical debt to avoid unexpected hurdles during the migration process.
While the blog post raised alarms regarding IBM’s mainframe business, it came three years after IBM had recommended using AI for COBOL migration. The company even launched the “watsonx Code Assistant for Z” to facilitate this process. This past month, IBM reported its highest mainframe revenue in two decades, attributing this growth partly to its AI-driven code conversion tools. CEO Arvind Krishna stated that mainframes continue to provide the most cost-effective solution for specific workloads, despite the challenges posed by legacy systems like COBOL.
The pressure to modernize COBOL applications is not a new concern. Just last week, Infosys Chairman Nandan Nilekani noted that advancements in AI technology have made the cost of rewriting legacy applications more affordable and necessary. Recent initiatives from firms like AWS, Microsoft, IBM’s spin-out Kyndryl, and NTT have further highlighted the ongoing discussions around mainframe migration.
Despite these efforts, IBM’s stock plunge also coincided with rising unease within the SaaS industry, as speculation grew that AI could disrupt established business models. Companies such as Salesforce, Atlassian, Adobe, ServiceNow, and HubSpot have all seen significant drops in their share prices amid these concerns.
Anthropic’s blog post underscores the enduring risks associated with legacy technologies like COBOL and illustrates how AI can re-energize discussions around these issues. The commentary reflects a broader conversation in the tech community about the balance between legacy system challenges and the potential benefits of AI-driven solutions.
