The 30 second story
Picture buying a house based on glossy estate agent photos, only to discover the roof leaks and the boiler is broken. Gartner has found that AI-powered projects to move businesses off old mainframe computers are failing far more often than the sales pitch suggests. The research firm analysed dozens of these migration projects and found that most companies either abandon them halfway through or end up spending far more than budgeted while getting less than promised.
Why it matters
Mainframes are the massive, expensive computers that run the core systems for banks, insurers, and large manufacturers. Moving off them has always been attractive because mainframes cost a fortune to maintain and require specialist staff who are retiring faster than they can be replaced. The promise of AI tools that can automatically convert old mainframe code to modern systems sounded like the perfect solution. But Gartner’s analysis shows these AI tools struggle with the complexity of systems that have been built up over decades. Companies are finding themselves trapped in expensive, half-finished projects with systems that work worse than what they started with. The automation that was supposed to make these transitions smooth and predictable is not mature enough to handle the job reliably.
What this means for your business
- Mainframe-dependent suppliers may face higher costs and service disruptions as their migration projects stall or fail
- The skills shortage for maintaining old systems will get worse as more migration attempts fail and companies stick with mainframes longer
- AI automation works best on newer, simpler tasks rather than complex legacy system overhauls
- Technology transitions that sound revolutionary often cost more and deliver less than traditional approaches