IBM’s stock experienced one of its biggest single-day drops on February 24, in over 25 years, due to market concerns about new AI tools (like Anthropic’s Claude Code) that claim they can automate the modernisation of COBOL code. The stocks which fell 13 percent made the investors worried that if COBOL (programming language) becomes easier to update or replace with AI tools, IBM’s legacy mainframe services could lose revenue. Mainframes, which run critical COBOL systems, are a major source of revenue and differentiation for IBM, and hence any fear that this could be disrupted sent ripples through the stock price.What is COBOL?COBOL stands for Common Business-Oriented Language. It’s one of the earliest programming languages ever created, dating back to 1959, and was designed specifically for business and administrative computing. It was built so computers could efficiently handle things like payroll, banking, inventory and other large business tasks. It uses readable, almost English-like sentences (for example, ADD/SUBSTRACT amount) so humans can understand the code easily. Built in the US, a lot of big banks, governments, insurance companies and large enterprises started adding it to their mainframes (high-performance, ultra-reliable computers designed for large-scale, secure, real-time data processing and transactions) and still run critical systems written in COBOL on them. Despite being old, COBOL still processes huge portions of global financial transaction volume because those systems are extremely stable and trusted.COBOL is also a key technology tool for many Indian IT firms, especially those serving global banking, insurance and retail infrastructure. Tata Consultancy Services (TCS) is one of India’s largest employers for mainframe and COBOL roles, largely supporting international banks. Infosys and Wipro also hire developers and use COBOL across multiple IT and consulting projects.What is the concern?According to IBM, an estimated 250 billion lines of COBOL code are in production use across sectors such as financial services, government, logistics, manufacturing and retail. COBOL supports 80 percent of in-person credit card transactions and 95 percent of all ATM transactions in the US, while generating over $3 billion in commerce each day. The main concern is that COBOL is old, and many programmers trained in the language have retired. There’s a shortage of people who can maintain these systems. Hence, many organisations are slowly modernising or rewriting COBOL code in newer languages. IBM engineers themselves have raised concerns about the same.Anthropic entered the picture when it shared a blog post arguing that modernising COBOL manually is becoming difficult day by day and Anthropic's AI tool Claude Code could automate much of the complex analysis work involved in modernising COBOL. The post said: “The developers who built these systems retired years ago, and the institutional knowledge they carried left with them. Production code has been modified repeatedly over decades, but the documentation hasn't kept up. Meanwhile, we aren't exactly minting replacements—COBOL is taught at only a handful of universities, and finding engineers who can read it gets harder every quarter.”It further mentions that modernising a COBOL system required armies of consultants spending years mapping workflows which resulted in large timelines and high costs that only a few were willing to take on.How AI can change COBOL modernisationAccording to Anthropic, here's how AI can help:Full codebase mapping: Anthropic says AI can read entire COBOL systems, trace execution paths, map data flows and uncover hidden dependencies that make modernisation risky.Risk-based assessment: It can identify tightly coupled modules, isolated components, duplicated logic and technical debt to prioritise safer, high-value upgrades.Smarter planning: AI can recommend data sequencing based on risk and complexity, while human teams set business priorities, compliance needs and target architecture.Stocks plummeting and IBM’s reactionShortly after Anthropic's blog post, IBM suffered its worst market day, with shares plunging 13 percent. According to Bloomberg data, the stock is down 27 percent in February, on course for its steepest monthly decline since at least 1968. The selloff was driven by investor concerns that AI could automate work once dependent on years of consulting, potentially hurting IBM’s revenue streams.The fear spread to Indian IT stocks as well. Coforge, Persistent Systems and HCLTech fell by 7 to 8 percent, while Infosys, Tech Mahindra, Mphasis and TCS declined by 4 to 6 percent. The Nifty IT index dropped 6 percent on Tuesday.Defending IBM's position, responding in a February 23 blog post, IBM’s Senior Vice President and Chief Commercial Officer Rob Thomas said: “The value IBM mainframe delivers has nothing to do with COBOL. It has to do with what the platform is: A purpose-built architecture from silicon through the operating system for unmatched transactional resilience, security, performance and efficiency at scale that no other distributed environment has been able to deliver.”