Anthropic’s announcement that its updated AI tool can modernise COBOL, an old programming language that runs on IBM’s mainframes, left Big Blue’s stocks battered on Monday. Banasree Purkayastha looks at how the AI startup’s tool aims to automate the legacy system which is scaring the enterprise software and IT services industry
l What did Anthropic say?
IN A BLOG post on February 23, AI startup Anthropic declared that the economics of modernising COBOL, a dated programming language that runs on IBM mainframes, have shifted with the advent of AI tools like its own Claude Code. “Modernising a COBOL system once required armies of consultants spending years mapping workflows.
This resulted in large timelines and high costs that few were willing to take on. AI changes this,” it said. “With AI, teams can modernise their COBOL codebase in quarters instead of years.” Investors saw it as a potential threat to IBM’s business model. Big Blue’s stocks plunged 13.2%, its biggest single-day drop since October 18, 2000, wiping $40 billion off the company’s market value.
l What is COBOL?
COBOL OR COMMON Business-Oriented Language, is one of the oldest programming languages which is still widely used for business, financial and administrative data processing. It serves as the foundation for more than 40% of all online banking systems, supports 80% of in-person credit card transactions, handles 95% of all ATM transactions, and powers systems that generate more than $3 billion of commerce each day. It was created in 1959 by a consortium called the Conference on Data Systems Languages. The English-like compiled programming language enables programmers to use a readable, easily maintainable programming language that can function across mainframe computers and operating systems.
l Why modernising it is a challenge
AS A LEGACY programming language, COBOL is still deeply embedded in global enterprise infrastructure. However, COBOL modernisation is required for moving mainframe workloads to the cloud. While modernising COBOL has been a technically solved problem for a while, it is the high cost of modernisation that’s the real stumbling block. It requires reverse-engineering business logic from systems built decades ago and there are very few engineers around who can read COBOL.
l What Claude Code hopes to change
ANTHROPIC SAYS AI tools can accelerate refactoring of apps written in the COBOL language. Making a Code Modernisation Playbook available for download, it said Claude Code can automate the exploration and analysis phases that consume most of the effort in COBOL modernisation. It said AI translates COBOL logic into modern languages, creates API wrappers around legacy components that stay in place, and builds the scaffolding to run old and new code side by side during transition.
Claude Code can analyse codebases, map hidden dependencies, and generate working translations of code that most engineers today cannot read. For enterprises that run COBOL on distributed platforms such as Windows or Linux, modernisation of the legacy system could be useful since Claude Code would enter a space where IBM’s vertical integration may not offer much benefit.
l IBM’s response to the challenge
IBM OWNS THE mainframe platform on which the code runs and thus earns recurring revenue from COBOL workloads. Governments, insurance companies, banks and others pay IBM for mainframe hardware refresh cycles, software licenses and performance upgrades. The technology giant has said its core mainframe computer business
offers a platform that provides the same quality of performance and security for various programming languages and not just COBOL. In an interview to a business daily, IBM chairman, president & CEO Arvind Krishna said the scrip fall was an overreaction. He said that most legacy systems run on mainframes because they are better equipped to handle such workloads.
In 2023, IBM had introduced its own AI-powered mainframe coding assistant — watsonx Code Assistant for Z — to help migrate COBOL to modern languages such as Java.
l Worrying signals for Indian IT industry
INDIAN IT STOCKS extended their slide on Tuesday, with the NSE Nifty IT Index losing nearly 5%. The trigger, say experts, is what the technology signals – AI may sharply compress the cost and timelines of large legacy modernisation projects — work that has been a steady revenue stream for Indian IT services firms. Srikanth Velamakanni, co-founder of analytics firm Fractal, said AI could reduce coding costs from roughly $15 per line to about $2, thus shrinking a billion-dollar transformation into a $100-million exercise.
“There are roughly 800 billion lines of code globally that can be modernised. Even at $2 per line, that becomes a $1.6 trillion opportunity. AI compresses everything, but it exposes the entire installed base to modernisation,” he explained. A report by Citrini Research added to the fears as it painted a scenario where AI agents cause mass unemployment for white-collar workers, shrink consumer spending and induce economic contraction, as early as 2028.