Shares of IBM experienced a significant drop, the largest in over 25 years, due to growing apprehensions regarding the impact of artificial intelligence on one of the company’s core operations. The stock tumbled by 13.2%, closing at $223.35, marking its most substantial single-day decline since October 18, 2000.
Investors have reevaluated their stance on IBM, with shares plummeting approximately 25% this year, amid concerns about the rapid transformation artificial intelligence could bring to the landscape of enterprise software and IT services. The recent sell-off was instigated by a blog post from AI startup Anthropic, which introduced its AI tool, Claude Code, capable of comprehending and modernizing COBOL, an aging programming language pivotal to IBM’s mainframe business.
COBOL, a language originating in the 1950s, remains extensively utilized in crucial sectors such as banking, aviation, insurance, and government agencies, playing a vital role in IBM’s revenue stream. The process of updating COBOL systems has historically been sluggish, costly, and reliant on substantial consultant teams, providing a steady income source for IBM.
Anthropic’s assertion is that AI revolutionizes this dynamic by simplifying the analysis and enhancement of legacy code. The company highlighted that despite the declining number of COBOL experts, there are still hundreds of billions of lines of COBOL code operational daily in live systems.
AI’s proficiency in managing intricate and time-consuming tasks that previously made COBOL system modernization economically unfeasible is emphasized by Anthropic. Notably, around 95% of ATM transactions in the US are reportedly reliant on COBOL, underscoring the language’s deep integration within financial infrastructure.
Anthropic’s AI tool boasts the capability to scan extensive codebases, map software interdependencies, generate lucid system documentation, and identify potential risks that would conventionally require months to uncover manually. The company contends that the cost barrier associated with comprehending outdated code surpasses that of rewriting it, a paradigm shift facilitated by AI.
