AI adoption is accelerating among mainframe users, and it’s spurring increased usage of the big iron platform, according to Kyndryl’s annual State of Mainframe Modernization Survey.
Most mainframe customers are implementing generative AI on the mainframe, and they’re expecting their AI investments to drive significant cost savings and boost revenues, new research finds.
Collectively, mainframe customers surveyed by Kyndryl are forecasting $12.7 billion in cost savings and $19.5 billion in increased revenues. Among the organizations polled, nearly 90% have implemented or plan to implement generative AI tools in their mainframe environments. Top use cases include enhanced performance and resource allocation, fraud detection, and security and threat detection, according to Kyndryl’s annual State of Mainframe Modernization Survey. AI is also helping mitigate skills gaps across the workforce, Kyndryl reports.
In the big picture, more than half of respondents (56%) told Kyndryl they’re increasing their use of mainframes and leveraging them in new ways within their hybrid IT environments. Drivers include new technologies and approaches such as agentic AI, LLMs and DevSecOps.
“Last year, over a third of enterprises cited opportunities offered by AI, including agentic and generative AI, as a key influence on their mainframe modernization investment decisions. In 2025, confidence in the technology continues to grow. When deployed on the mainframe, AI and generative AI can help derive powerful insights from complex data, convert code, modernize applications and increase workforce efficiency,” Kyndryl stated.
AI adoption is moving quickly, Kyndryl noted. Most organizations (80%) were still in the early or mid-stages of AI adoption in 2024, but now progress is accelerating. “15% of respondents have now completed or are nearing the final stages of AI integration versus just 4% last year. This momentum reflects not just growing technical maturity but also rising confidence in AI’s potential to deliver measurable business value,” Kyndryl stated.
“Organizations are taking a much more pragmatic approach to modernization to drive better business outcomes and incorporate new technologies,” said Hassan Zamat, global practice leader for core enterprise at Kyndryl, in a statement.
Another significant survey finding is that mainframe modernization plans are more dynamic than they’ve been in the past due to changing market conditions: 80% of organizations reported that their modernization strategy has shifted in the past year, “indicating widespread re-evaluation and a degree of agility not commonly associated with technology projects of such significant size, scope and complexity,” Kyndryl stated.
Mainframe modernization can take a few forms; enterprise customers can keep data on the mainframe, migrate workloads off, or create a hybrid environment.
Among organizations that said they’ve shifted their modernization approach, 43% are more focused on modernizing directly on the mainframe, and 50% are moving more towards a hybrid strategy, Kyndryl found. Among those that are moving towards a hybrid IT approach, 34% are prioritizing integration with cloud platforms, and 16% are increasing the pace of moving applications off the mainframe. Only one out of 500 respondents reported plans to move entirely off the mainframe.
The reasons for changing their migration strategies vary widely among respondents and include factors such as the success of previous projects, budget shifting to new technologies, and external factors such as geopolitical, macroeconomic and regulatory concerns, Kyndryl stated.
Cost is another motivator: The average cost of modernizing on the mainframe dropped from $9.1 million in 2024 to $7.2 million in 2025, and the ROI on such a move more than doubled, Kyndryl stated.
Other trends noted in the mainframe modernization report include:
- AI isn’t for everyone: “Despite the possibilities, some enterprises remain reluctant to deploy these emerging technologies. Twelve percent are not yet planning to make use of AI and generative AI. The main barriers include security protocols that limit access to mainframe data (47%), the perception that generative AI technology lacks maturity (41%) and mainframe-specific technical barriers (40%),” Kyndryl stated.
- Multi-skilled organization are a must: “It is no longer enough for organizations to have experts with skills in mainframe technology only. As mainframes are increasingly operated within hybrid environments, organizations also need experts with knowledge of cloud, security, AI and generative AI, as well as networking and edge computing. Seventy percent of respondents have had difficulties finding the right skills, and three quarters of respondents (74%) leverage external firms to help them modernize.”
- Compliance requirements are influencing modernization plans: “Ninety-four percent of respondents report that regulatory compliance strongly influences their organization’s decision-making regarding mainframe modernization, causing organizations to accelerate timelines, increase collaboration across functions and design for compliance from the outset. Furthermore, 32% cite security as a reason why they kept an application on the mainframe.”
Kyndryl commissioned Coleman Parkes Research for its study, which surveyed 500 senior leaders at mainframe-based enterprises with an average global revenue of $3.6 billion.
