The Immortal Legacy
COBOL has been declared dead for four decades. Every generation of developers inherits the same prophecy: modern languages will finally replace it, legacy systems will be rewritten, and COBOL's extinction is inevitable. In 2018, TSB Bank tested this theory with a migration away from their COBOL infrastructure and the result was catastrophic. The project locked 1.9 million customers out of their accounts, cost over £330 million, and ended with the CEO's resignation. The incident proved what many suspected: for systems processing trillions in transactions annually, migration risk often exceeds the cost of maintenance. When LLMs arrived in 2023-2024 promising automated code translation, it seemed like COBOL's death warrant had finally arrived. Instead, vendors began pitching "AI-powered COBOL modernization", which still meant running COBOL, just containerized in cloud infrastructure. The pattern became undeniable: every solution to COBOL becomes another reason to keep using COBOL.
Evolution, Not Revolution
After TSB's disaster demonstrated the existential risk of replacement, major institutions made a calculated pivot: modernize the infrastructure, keep the language. Engineers at Citi published compiler upgrade strategies focused on performance optimization and "future-ready features." Technical blogs detailed containerizing COBOL applications with Docker for cloud-native deployment. AWS published guides for running COBOL workloads on EKS with generative AI assistance. Not to replace COBOL, but to integrate it with modern tooling. Consulting firms like Cognizant hired senior architects specializing in Raincode transformations, acknowledging that COBOL modernization requires deep expertise, not automated shortcuts. The approach is consistent across the industry: everything modernizing COBOL today is built around it, not instead of it. These aren't exit strategies; they're capital expenditures that extend COBOL's operational horizon. The industry won't admit the uncomfortable truth: COBOL isn't just resilient, it's economically rational.
The Language That Refuses to Die
In 2025, Over 30 major organizations are actively hiring COBOL developers, including financial institutions like Bank of America, Citi, Deutsche Bank, and UBS; insurance companies like Zurich and Unum; and unexpectedly, automotive companies like Volkswagen and Hyundai MOBIS. Volume leaders like Citi and Crédit Mutuel maintain substantial COBOL teams, while consultancies build specialized modernization practices. The hiring market reflects a simple truth: institutions don't invest in skills for technologies they expect to abandon. Far from declining, COBOL hiring has remained structurally resilient through multiple technological disruption cycles: Y2K remediation, offshore outsourcing waves, the "rewrite everything" movement, and now the LLM revolution. Every supposed solution becomes another infrastructure layer supporting COBOL's continued operation. The language isn't dying; it's being modernized, containerized, and architect-designed into the next decade. COBOL has had its funeral scheduled for forty years. It has yet to attend.