Anthropic's positioning of its Claude Code AI model as a solution for modernizing legacy COBOL (Common Business-Oriented Language) code is sending ripples through the tech world, particularly impacting IBM. This matters because it challenges IBM's existing strategy of keeping modernized COBOL code running on its own mainframes, potentially disrupting a key revenue stream. The market's reaction, including a selloff in software stocks, shows the perceived threat to established software business models [1].
COBOL, a programming language widely used in business applications, is facing a talent shortage as experienced programmers retire. The language, often running on mainframes, has become increasingly complex over decades of modifications. Anthropic is betting that Claude Code can help organizations migrate this code to modern languages and cloud environments, reducing their reliance on aging infrastructure and scarce COBOL expertise.
IBM's strategy involves offering AI tools, such as watsonx, to modernize COBOL code. However, a key difference is that IBM aims to keep the modernized code running on its mainframes. "The strength of our Z placement fuels our flywheel for growth with its attractive 3x to 4x stack multiplier across IBM," said CFO James Kavanaugh after its latest earnings report. This approach allows IBM to maintain control over its ecosystem and continue generating revenue from its hardware and software offerings.
Anthropic argues that Claude Code can automate COBOL modernization and incrementally migrate code into modern languages, where it can be hosted across various cloud providers. This offers organizations greater flexibility and potentially lower costs compared to staying locked into a single mainframe architecture. COBOL is standard in many financial operations (like ATMs), as well as in government and airline systems, as Anthropic notes, so users may want to keep this code tied to one mainframe architecture for security, reliability, and speed (it’s the devil they know!) rather than migrating to a different platform.
The implications extend beyond COBOL modernization. Anthropic's release of Claude Code Security, a tool for finding software vulnerabilities, has also rattled the cybersecurity industry. This tool "reasons about your code the way a human security researcher would," potentially offering a more effective alternative to traditional static analysis tools, according to industry experts. This has led to a decline in shares of several major cybersecurity providers.
Figma's partnership with Anthropic further illustrates the potential for AI to disrupt software development. The "Code to Canvas" feature converts AI-generated code into editable designs within Figma, creating a bridge between AI coding tools and the design process [2]. If AI tools keep improving, teams may eventually skip the design refinement step altogether.
More insights on trending topics and technology
