Gartner warned on June 18 that more than 70 percent of mainframe migration projects launched this year will fail to deliver targeted results because organisations overestimate the capabilities of generative AI tools.

Gartner also forecast that 75 percent of vendors active in the mainframe migration market will change their business models or cease operations by 2030.

It attributed that to market expectations being tempered and demand for uniform migration solutions declining.

Gartner Vice President Alessandro Galimberti (알레산드로 갈림베르티) pointed to a widening gap between generative AI marketing messages and the actual ability to transform and migrate complex legacy code. He said investor pressure is pushing vendors to force AI into products regardless of performance. He added that as the “too big to fail” nature of mission-critical mainframe applications combines with an exodus of skilled personnel, unstructured migration strategies will become an increasingly hard-to-manage risk.

Gartner recommended different strategies by organisational size. It said mid-sized environments should focus on optimising existing mainframe investments and apply full platform migration selectively. It said smaller environments should consider mainframe-as-a-service as a cost-effective operating model and focus on replacing legacy software and modernising within the platform.

Galimberti said that for many mainframe customers, generative AI can be used more effectively not as a tool to speed platform migration but as a tool to support modernisation within the platform.