ISG Provider Lens™ Mainframes - Services and Solutions - Mainframe Operations - U.S. 2025

The year GenAI transformed mainframe modernization and extended mainframe future

Generative AI (GenAI) has a dual effect in the mainframe modernization market. It accelerates migrations off the mainframe and provides clients with tools to extend their use of mainframes.

In the Mainframe Optimization Services quadrant, ISG observed a growing trend among clients to adopt DevOps, containers and APIs to enhance their application development and maintenance platforms. These clients recognize that their mainframes will continue operating for many more years, making investments in modernizing tools and processes essential. A significant demand many service providers report is integrating mainframe data with the public cloud. Changes in market regulations have also created new demands around security and compliance. Enterprises must update their legacy applications to meet these new requirements, and providers use GenAI’s code inspection functionality to accelerate the discovery of changes and the modernization of legacy code.

However, to use DevOps, containers and GenAI with legacy languages, such as COBOL, enterprises must update their application development workbench, including testing automation, code repositories and version control. Once these functions are enhanced, enterprises realize their mainframes can exist longer, providing agile services and rapidly creating APIs to meet data integration demands.

In the Application Modernization Services quadrant for the U.S., ISG observed that GenAI has opened new possibilities and elevated clients’ trust in application modernization. Companies facing rising mainframe costs prefer replatforming their mainframes to the cloud to achieve savings quickly. However, GenAI has enhanced reengineering and rewriting processes with rich documentation and extended testing while reducing the time required to complete modernizations. This study finds that companies’ preferences have shifted from replatforming to reengineering methods, and the new demand comes with additional requirements. Until last year, typical clients expected providers to ensure full equivalence — a like-for-like behavior and outcomes — between the legacy and new applications. Today, more clients anticipate incorporating innovations into the new code to add business value to the transformation.

Providers can respond to these requests with GenAI, allowing clients to validate the assessment documentation and insert new requirements with GenAI, which checks dependencies and impact. A safe alternative is identifying  intersection points to create APIs and microservices that add the required innovation.

GenAI can inspect code and explain how it works, reducing and sometimes eliminating the dependence on senior experts’ knowledge, thus reducing the cost of maintaining legacy applications.

The Application Modernization Services quadrant for Brazil shows the same challenges and demands as the U.S.  quadrant. Companies in Brazil are demanding more modernization, with new client cases emerging in the banking, insurance and retail sectors. In December 2024, Banco Itaú Unibanco, the largest retail bank in Latin America, announced at AWS re:Invent in Las Vegas its plan to complete its migration to AWS by the end of 2028. The bank serves more than 90 million clients and previously had the largest mainframe in Brazil, with over 300,000 MIPS just a few years ago. Many service providers in this quadrant are collaborating with Itaú Unibanco, and other large banks are also pursuing cloud migration. Open banking and the instant payment system Pix, operating in the cloud, have transformed the competitive landscape, along with digital banks running on AWS, such as Nubank and C6 Bank, with 100 million clients and 30 million clients, respectively.

The Mainframe as a Service (MFaaS) quadrant shows annual market growth of 5 percent. Most new deals are from clients who previously outsourced their mainframes and are seeking to modernize to a pay-as-you-go model, thereby freeing up capital spending for investment in other business opportunities. Another group of clients comprises companies that are outsourcing for the first time. The MFaaS is a stable market; it is uncommon to see clients switching providers. New clients should assess the level of automation offered and the speed at which providers  upgrade their hardware and software to prevent obsolescence. Clients using obsolete technologies must agree to  upgrade their systems to fully benefit from MFaaS.

New trends in the Mainframe Operations quadrant include replacing middleware to eliminate costly licenses and optimizing infrastructure based on industry standards and best practices. Custom solutions have become less viable as mainframes no longer provide business differentiation. Service providers are investing in automation and GenAI to improve CX and rapidly identify optimization opportunities.

While providers prepare for new IBM solutions related to AI and GenAI on mainframes, they report a lack of client demand for running AI on these systems. More common requests involve APIs and data virtualization services to  integrate mainframes with cloud-native applications. The true value of GenAI in operations lies in its capacity to inspect applications for code-optimization opportunities, which can reduce MIPS consumption and improve application
performance. Providers frequently encounter dead code, duplication and old programming methods that consume mainframe power without adding value, and GenAI quickly identifies these issues.

A new challenge in mainframe operations is related to security, as some providers reported that their clients fell victim to ransomware attacks, driving increased demand for security services and immutable storage solutions.

In the Mainframe Application Modernization Software quadrant, the major change is GenAI. Most vendors have adopted GenAI to enhance their modernization offerings. However, GenAI and conventional large language models  (LLMs) have yet to fully grasp application architecture and software engineering to provide guidance on designing software architecture. Vendors in this quadrant offer robust toolsets to assess applications, map dependencies and refactor code bases. Writing new code is faster with these tools based on software engineering methods.

GenAI addresses several gaps in the modernization process, accelerating and mitigating risks associated with  mainframe modernizations. In most mainframe migrations to the cloud, more than 70 percent of the time is spent on testing. GenAI significantly accelerates this process by generating testing scripts from the legacy source code and
automatically running them on the newly created applications, ensuring speed and accuracy. In addition to  accelerating testing, GenAI can pinpoint bugs in the code, facilitating quick fixing and improving quality. GenAI is also
used as a human interface, enabling clients to inquire about the functionality of both old and new code, understand business rules and assess the impact of modifications. Other reported benefits include eliminating dead code, replacing code libraries and providing clean code for future maintenance.

Access to the full report requires a subscription to ISG Research. Please contact us for subscription inquiries.