The 2026 BMC Mainframe Survey, which gathered data from over 1,300 global decision-makers, reveals a clear preference for "cautious realism." While 45% of respondents identify AI implementation as a top priority, they remain hesitant to grant the technology full autonomy. Currently, only 23% of practitioners are comfortable allowing AI to execute code-management tasks independently, even though 40% are willing to let it provide recommendations. A similar pattern emerges in database management, where the gap between seeking advice and authorizing action remains significant.

Operational concerns continue to shape investment strategies. High implementation costs, cited by 41% of respondents, and security risks weigh heavily on deployment timelines. Furthermore, the industry faces an impending technical hurdle regarding digital certificates. With upcoming 2029 requirements to shorten TLS certificate lifecycles, many organizations remain vulnerable; currently, one-quarter of mainframe shops still rely on manual processes for certificate management. To address these pressures, 36% of firms plan to invest in proprietary agentic management, while 32% look toward third-party solutions to maintain operational resilience.