The exit narrative has serious engineers behind it now. We are building for the enterprises that stay.
There is a well-funded, serious effort underway to help enterprises exit mainframe.
The tools are genuinely impressive. The engineering is real. The problem they are solving is real.
For years, mainframe modernization was sold by consultants who had never seen a 3270 datastream. The pitch was simple: COBOL is old, cloud is new, we will migrate everything and your problems will be solved.
The results were consistent. A major European retail bank that missed its batch window by twenty-two hours after moving to cloud. A US financial services firm that burned tens of millions of dollars before rolling back to the mainframe. Countless unnamed projects that died quietly in budget reviews three years in.
Something changed recently. The modernization space attracted serious engineers. People who understand that a translated program can compile and pass tests while still behaving differently from the legacy system under real operating conditions. People building verification loops, simulation twins, and comparator harnesses.
This is meaningfully better than the previous generation of migration promises.
Gartner published their State of the IBM Mainframe report this year. A notable finding: for some organizations running VMware workloads, mainframe can be cheaper than Broadcom’s new licensing. The analysts noted this with some surprise.
They should not have been surprised. The math has always been there.
IBM reported the highest IBM Z revenue in twenty years at Think 2026. The z17 launch was a record. These are not organizations preparing to exit. These are organizations investing in the platform.
$10 trillion in daily transactions runs on z/OS. 95% of the world’s top banks run mainframe.
The exit market is real. Some organizations will modernize. Some already have. But the stay market is larger. It is stickier. And it has almost entirely different needs from the exit market.
An organization that has decided to stay on mainframe is not standing still. It is facing real operational challenges.
Security requirements that are getting stricter, not looser. Audit demands that are getting more specific, not less. A skills gap that is getting wider, not narrower. AI capabilities arriving on the platform that need to be integrated, not ignored.
These organizations do not need tools designed to translate them away from what they run. They need tools designed for the reality they operate in.
The gap between what those organizations need and what the current wave of mainframe AI tooling provides is significant.
That gap is what we are working on.
Better tools for the wrong destination are still the wrong tools.
The organizations that have decided to stay deserve tools built for the platform they actually run.
That is what we are building.
Related reading: People Can Sleep Peacefully With Mainframe · Why AI Pilots Succeed and Productions Fail · What IBM Think 2026 Got Right – And What It Left Out