On Tuesday, Anthropic published tools that let the cloud read, analyze, and translate legacy COBOL into modern languages like Java and Python. By the end of the trading day, investors had wiped out nearly $40 billion from IBM’s market cap – the company’s largest single-day decline in 25 years – appraising the announcement as a potential threat to IBM’s mainframe business.
The reaction was intense. It was also built on a fundamental misconception of why enterprises run mainframes in the first place.
IBM’s COBOL is 66 years old. It was designed in 1959, runs on IBM mainframes, and continues to power transaction processing systems with an estimated 250 billion lines of COBOL in active production according to the Open Mainframe Project.
The engineers who wrote it are retiring; The people replacing them may not read it at large. For decades, the skills gap has been one of enterprise IT’s most expensive unsolved problems — and one IBM has been working to fix with AI since at least 2023, when it launched the Watsonx Code Assistant for Z to help migrate COBOL to modern Java.
Anthropic says Cloud Code can now analyze entire codebases, map hidden dependencies and generate working translations of code that most engineers can’t read today. For enterprises running COBOL on distributed platforms – Windows, Linux, and other non-mainframe environments – this capability is really useful and increasingly practical.
The real obstacle was never technical
"Modernizing COBOL has been a technically solved problem for some time," Gartner analyst Matt Brasier told VentureBeat. "The real problem is that the cost of modernization is high and the ROI is low."
Amazon and Google have been offering AI-powered COBOL migration tools for years. AWS Transform and a comparable Google Cloud Platform service both targeted the same problem: reducing friction for customers wanting to move mainframe workloads to the cloud.
"This is basically another source of competition," Raj Joshi, senior vice president at Moody’s Ratings, told VentureBeat. "IBM has always been in a very competitive field. At the margins, this thing is fundamentally negative, there’s no question about that. There is a more powerful competitor. But IBM is able to co-exist with these threats."
Steve McDowell, chief analyst at NAND Research, cuts to the structural argument: "The applications do not run on mainframes because they are written in COBOL," He said. "They run on mainframes because mainframes provide a range of determinism, scalable computation, and reliability that general purpose servers cannot match."
The issue goes deeper than just the state of the market. "GenAI tools are helpful, but their non-deterministic nature means that the resulting code is not consistent – the same operation will be applied in different ways in different parts of the code," Brassier said. "Leading tools combine deterministic and non-deterministic approaches. However, none of these solve the ROI problem."
What does COBOL translation leave unresolved?
"Translating COBOL is the easy part," IBM communications director Steven Tomasko told VentureBeat. "The real work is data architecture redesign, runtime replacement, transaction processing integrity, and hardware-accelerated performance built over decades of tight software and hardware coupling. This is the problem IBM has spent decades trying to solve, and AI is the most powerful tool we have yet to solve it."
According to IBM, the Royal Bank of Canada, the National Organization for Social Insurance, and ANZ Bank have used Watson’s Code Assistant for Z to accelerate the modernization of COBOL code without moving from IBM Z.
This does not mean that Anthropic has no competitive edge. For enterprises running COBOL outside the mainframe – on distributed systems, Windows and Linux environments – cloud code enters a space where IBM’s vertical integration is nothing short of an advantage. "IBM understands mainframe technology at a level that others cannot match. If I’m only looking at COBOL, I’m using IBM’s watsonx," McDowell said. "However, Anthropic has a broad presence across multiple development teams, where a single vendor makes it worthwhile."
What business buyers should really do
Senior data and infrastructure engineers will spend the next few weeks asking questions of executives who saw the headlines and assumed the hard problem had been solved. I didn’t.
"This is COBOL, but there are many applications related to it," Joshi said. "It’s not like you change millions of rows and somehow you’re ready to go to the cloud. It’s a massive assessment of risk, dependency and all those things."
The more useful question for buyers is whether this week’s noise creates an opening. Brazzers thinks it does.
"They should use the resulting board-level and shareholder discussions to review deferred modernization initiatives and see if any of them now have an ROI," Brassier said.
McDowell was blunt on the competition question. "Will Anthropic take business from IBM’s tools? yes of course," He said. "But I would be surprised if that device is generating significant revenue for IBM."
Constellation Research analyst Chirag Mehta cautioned that IT leaders should not react emotionally or rewrite strategy overnight.
"Consider this as a reason to run a small, limited pilot to measure results, not as a reason to break up and change vendors." Mehta told VentureBeat.
Mehta suggests that enterprises choose a well-scoped application slice or workflow with clear inputs and outputs, and take an apple-to-apple approach to evaluate: the quality of dependency mapping, the quality of retrieved business logic documentation, test coverage and peer checking, performance and reliability regressions.
In Mehta’s view, the big reminder is that modernization is about more than changing code. The hard parts are extracting institutional knowledge, reworking processes and controls, making change management, and incorporating operational risk into systems that can’t be broken. AI may compress the “analysis and translation” work, but it does not eliminate the burden of governance and accountability.
"Winning teams will treat AI as an accelerator inside a disciplined modernization program, with measurable checkpoints and risk guardrails, not as a magical conversion button." Mehta said.
<a href