In a stunning reminder that artificial intelligence is rewriting the rules of the tech industry in real time, shares of International Business Machines Corp. plunged 13% Monday — their steepest single-day drop in more than 25 years — after an AI startup suggested it could dismantle one of IBM’s most durable competitive moats.
The trigger? A new coding tool from Anthropic that promises to modernize decades-old COBOL systems — the very workloads that have kept IBM’s mainframe ecosystem entrenched inside banks, governments, and Fortune 500 companies.
🤖 The Tool That Shook a 113-Year-Old Giant
Anthropic unveiled enhancements to its Claude platform, including a feature dubbed Claude Code, designed to analyze and help migrate legacy COBOL software — a programming language dating back to 1959 but still deeply embedded in mission-critical infrastructure.
The company said what once required “armies of consultants spending years mapping workflows” can now be dramatically accelerated through AI-driven code analysis.
That message rattled investors who fear generative AI could compress one of enterprise tech’s most lucrative and slow-moving businesses: legacy modernization.
💻 Why COBOL Still Matters — And Why AI Is a Threat
Despite its age, COBOL continues to run enormous portions of the global financial system. Many of those workloads sit on IBM-built mainframes — specialized, high-reliability servers prized for security, uptime, and regulatory compliance.
A meaningful slice of IBM’s revenue still ties back to this ecosystem.
If AI tools can suddenly make it easier and cheaper to rewrite or migrate those systems, the economics of long-term maintenance — historically IBM’s stronghold — could change dramatically.
🛡️ IBM Pushes Back: “The Platform Is the Value”
IBM executives were quick to counter the narrative that its business depends on an aging language.
Senior Vice President Rob Thomas emphasized that the company’s mainframe value proposition lies in performance and security — not COBOL itself.
“Whether the application is written in COBOL, Java, or any other language, the platform provides the same guarantees,” Thomas wrote. “The language is not the source of that value. The platform is.”
In other words, IBM argues customers are buying resilience, not nostalgia.
📊 A Brutal Month for Software Stocks — Not Just IBM
Monday’s plunge capped a punishing stretch: IBM shares are now down roughly 27% in February, putting them on track for their worst monthly decline in decades.
The selloff reflects a broader reassessment of legacy software firms as AI coding tools rapidly evolve. Investors increasingly worry that so-called “vibe coding” — using AI to generate applications with minimal human input — could:
Reduce demand for traditional enterprise development tools
Shrink consulting-heavy modernization projects
Pressure pricing power across legacy vendors
AI releases from firms like OpenAI and Alphabet Inc. have intensified those fears, accelerating a rotation away from companies seen as vulnerable to automation-driven disruption.
🔍 Analysts Say the Threat May Be Overstated
Not everyone believes AI will trigger a mass exodus from IBM’s platforms.
Amit Daryanani of Evercore ISI noted that clients have long had the option to migrate away from mainframes — yet most have chosen not to.
That inertia reflects the enormous risk and cost associated with rewriting systems that handle trillions of dollars in daily transactions.
🧠 Ironically, IBM Has Been Using AI to Modernize COBOL Too
Lost in the market panic is the fact that IBM has already been applying AI to the same challenge.
The company introduced its own AI-assisted COBOL modernization tools in 2023, enabling developers to translate legacy applications into languages like Java.
CEO Arvind Krishna said in 2025 that adoption had been widespread, with many customers using AI not to abandon mainframes — but to better understand and selectively modernize them.
⚖️ The Real Battle: Speed vs. Stability
The episode underscores a fundamental clash shaping today’s enterprise tech landscape:
AI Startups | Legacy Platforms |
|---|---|
Promise rapid reinvention | Offer proven reliability |
Lower development barriers | Minimize operational risk |
Disrupt cost structures | Protect mission-critical uptime |
Banks and governments may be intrigued by AI-powered rewrites — but they remain wary of replacing systems that cannot fail.
🚨 A Symbolic Moment for the AI Economy
IBM’s sharp decline may ultimately be remembered less as a reaction to one product announcement and more as a psychological turning point.
For decades, legacy infrastructure companies were insulated by complexity and switching costs. AI is now testing whether that insulation still holds.
The question investors are suddenly asking is not whether modernization will happen — but how fast AI will force it.
For a company synonymous with enterprise stability, that uncertainty is something markets rarely price kindly.