Enterprise COBOL: the compiler that is not free.

IBM Enterprise COBOL for z/OS is the production COBOL compiler under most of the world's mainframe business logic: the core banking, insurance, and government systems that have run for decades and still process the transactions of record. It is the tool that builds and rebuilds those programs. The compiled code runs against the Language Environment runtime that ships inside z/OS, which is exactly why so many buyers underestimate the compiler as a licensed product in its own right.

Enterprise COBOL is licensed under the IBM International Program License Agreement (IPLA). You pay a one time charge for the perpetual license, typically priced in Value Units that scale with the capacity of the machine it is installed on, then an annual Subscription and Support (S&S) charge for fixes and new releases. It is not a Monthly License Charge product, so it does not flow through SCRT and the Rolling 4-Hour Average. A Value Unit Edition exists for qualifying new application workloads. The recurring exposure is the S&S stream, and that is where renewal attention belongs.

IPLA Value Unit pricing is tiered, not linear: the rate per unit usually steps down as capacity rises, so the licensed tier and the machine it sits on both shape the bill.

The dominant driver is the annual S&S on every licensed Value Unit position, which recurs whether or not active development is happening on that machine. The second driver is where the compiler is installed: because the license tracks the capacity of the host, putting a development compiler on a large production class machine prices it as if it were a production tool. The third is drift. Compiler entitlements left in place after a consolidation, a migration, or a development environment retirement keep generating S&S long after the work that justified them has gone.

The compiler versus runtime distinction is the trap that runs both ways. Common exposures we see at pattern level:

Enterprise COBOL renews on the S&S stream, so the levers that move it are about footprint and entitlement hygiene, not headline rate. The five that pay:

There is no realistic like for like swap for Enterprise COBOL in a production IBM Z shop. Open source COBOL implementations exist but do not match the production tooling, support model, and certified behavior an enterprise estate depends on, and recompiling a large body of business logic onto a different compiler is a migration program with real regression risk, not a renewal tactic. The credible moves are about discipline: consolidating where the compiler is licensed, qualifying workloads for Value Unit Edition where they genuinely fit, and pruning entitlements that no longer map to active development. Treat any pitch framing a quick compiler replacement as leverage with caution.