For many large organizations, phasing out mainframes is practically unfeasible. That is the conclusion of Gartner analysts in a recent report, which warns against overinflated expectations for generative AI in modernization projects.
According to the researchers, the role of AI in these types of projects is systematically overestimated. They write this in the report “Too Big to Fail: Why Mainframe Exit Projects Are Likely to Fail in the Age of Generative AI.” A large share of the initiatives launched in 2026 to phase out mainframes are not expected to deliver the results initially hoped for. The technology appears less capable of quickly and reliably replacing complex legacy environments than many vendors suggest, The Register reports.
AI as a measure of success
The current focus on AI plays a significant role in this, adds Heise. Within the investment world, an organization’s ability to apply AI is increasingly seen as a measure of success. This pressure is driving vendors to prominently position AI in a wide range of solutions, even when the added value is limited. According to Gartner, this leads companies to deploy AI in areas where it does not actually solve the problem.
In the case of mainframe migrations, this can have far-reaching consequences. Projects not only risk becoming more expensive than planned but also threaten the continuity of business processes. Errors in migration decisions can lead to prolonged disruptions and new technical debt, precisely in areas where stability is essential.
At the same time, there are also practical reasons why organizations are turning to AI. The shortage of specialists with mainframe expertise and the often substantial technical debt make modernization complex. Generative AI thus appears to be an attractive tool. In practice, however, the technology proves particularly suitable for providing insight into and documenting that technical debt, but much less so for automatically converting legacy code to modern platforms.
Market for mainframe exits shrinking
Gartner also expects the number of providers focused on supporting mainframe exits to decline sharply in the coming years. A large portion of this market will disappear or have to reorient itself because the promised results are not materializing, according to The Register.
Meanwhile, the mainframe’s position remains remarkably strong. Despite the rise of cloud-native architectures, the platform remains crucial for applications where reliability, security, and the processing of large volumes of transactions are paramount. Moreover, over the years, enormous amounts of data have been accumulated, which are tightly intertwined. For many companies, completely migrating these environments is simply not feasible, both technically and financially.
More and more organizations seem to be recognizing this. Instead of a complete exit, they are increasingly opting for a strategy focused on gradually improving existing systems. The ambition to completely phase out mainframes is thus diminishing, partly because the risk and costs of such projects are difficult to justify.