Research Finder
Find by Keyword
Kyndryl and Google Partner to Drive Mainframe Modernization.
Kyndryl and Google Cloud pair Gemini with mainframe expertise, aiming to accelerate modernization and data accessibility.
Key Highlights
- Kyndryl achieves Google Cloud Gemini specialization.
- New program targets streamlined mainframe modernization with generative AI.
- Focus on integrating mainframe data with Google Cloud's BigQuery.
- Generative AI to power code analysis and application rewriting.
The News:
Kyndryl and Google Cloud have expanded their partnership, leveraging Google's Gemini AI models to accelerate mainframe modernization. Kyndryl has achieved Google Cloud specialization for Gemini, and they are launching a new accelerator program. The collaboration aims to streamline application and data modernization. The joint effort seeks to enable customers to unlock mainframe data with generative AI. Find out more by reading the press release link.
Analyst Take
The collaboration between Kyndryl and Google Cloud to integrate generative AI into mainframe modernization represents a logical step in Kyndryl’s documented strategy to collaborate with the hyperscaler cloud provider, this time in the space where they have the largest concentration of client.
This initiative, rather than merely a conceptual leap, establishes a structured framework intended to expedite the intricate and often resource-intensive transition of mainframe workloads to cloud-based environments. Kyndryl has one of the most balanced approaches of any GSI to mainframe modernization, with a focus on modernizing on, with and off the mainframe. Also the MIPS under management by Kyndryl dwarfs that of any other GSI, with my estimates putting Kyndryl’s MIPS estate being six times larger than the next biggest players of Ensono and DXC who then distance the rest by some mark.
At its core, the partnership looks to leverage Google’s sophisticated Gemini models to dissect and refactor legacy mainframe code, aiming to simplify what has historically been a daunting process. This move is not without precedent, with the likes of AWS with Bedrock and Q taking the same approach amongst other solutions, but its significance lies in the deliberate pairing of Kyndryl’s extensive and global mainframe expertise with Google Cloud’s cutting-edge AI prowess, signaling a strategic intent to provide optionality for those workloads not suited to the mainframe.
The Kyndryl Mainframe Modernization with Gen AI Accelerator Program is engineered as a multi-phase endeavor, commencing with a comprehensive assessment and a strategic blueprint, followed by a meticulously guided modernization process. Clients might note that this phased methodology reflects an awareness of the complexities inherent in mainframe system overhauls, prioritizing clarity and manageability over rushed implementation. Put simply, the landscape of these projects is fraught with risk and strewn with failure, AI doesn’t radically change the risk profile.
Among the standout features is the integration of mainframe data into Google Cloud’s ecosystem, specifically platforms like BigQuery, Cloud Run, and Cloud SQL, which promises to unlock enhanced analytics capabilities and facilitate the training of AI models. This data liberation could prove transformative, particularly for industries reliant on historical data trapped in aging systems that have not been modernized in-situ leveraging newer mainframe technologies and sub-system enhancements..
The technological underpinnings of this initiative are robust, featuring tools such as the Mainframe Assessment Tool (MAT), Dual Run, and Mainframe Rewrite. Dual Run is the technology that Google attained through a collaboration with Santander back in 2022, which Google has been quiet about since the initial announcement of the collaboration, so I am interested to see how it has evolved since 2022 and will be looking for further briefings from the Google team now it is back in the spotlight.
These components collectively emphasize the role of generative AI in not only analyzing and documenting code but also rewriting applications to align with cloud-optimized architectures. The endgame appears to be the creation of technology stacks that are not just functional but optimized for the scalability and flexibility. Beyond the technical, the partnership extends into testing, certification, and risk mitigation, an acknowledgment that modernization is as much about reliability and compliance as it is about innovation.
Market dynamics further contextualize this announcement. The 2024 Kyndryl Mainframe Modernization Survey underscores a palpable demand for cloud migration and AI adoption, suggesting that the timing of this collaboration aligns with a broader industry shift. The ability to harness mainframe data for AI-driven applications emerges as a pivotal motivator, particularly as organizations seek to derive actionable insights from decades-old repositories. We will see IBM focus on this trend with the latest mainframe system launch slated for Q2 2025, so the trend is pervasive regardless of whether clients want to continue to stay current on mainframe technology or migrate off.
A cited example involving an insurance provider illustrates a practical application, offering a glimpse into how this technology might translate into real-world value. Observers will likely find it prudent to monitor adoption rates and the emergence of customer success narratives in the coming quarters, as these will serve as barometers of the program’s efficacy. I see a lot of positive sentiment for AI-driven code refactoring, especially of COBOL, but I am yet to see this materially affect the number of successful full scale net migrations off the mainframe. One successful COBOL application migration being a very different outcome that a wholescale decommissioning of a mainframe system
Looking Ahead
From an analytical perspective, the long-term success of this alliance hinges on its capacity to deliver measurable return on investment (ROI) and to demystify the modernization process for enterprises wary of disruption, risk and ultimately a failed project. A key trend to watch is the seamlessness with which customers can integrate their mainframe data into Google Cloud’s AI and analytics suite in a project timeline that aligns with the original business case. The potential to generate dynamic reports, train predictive models, and develop agile solutions could redefine operational efficiency for participants.
What is without question is that regardless of whether you firmly believe in the fundamentals of the mainframe or the cloud, this announcement reflects a growing industry consensus that legacy systems must evolve to meet the demands of an AI-centric future, a recognition that resonates across sectors.
For regulated industries, such as finance and healthcare, the ability to transform COBOL into Java while ensuring data residency and security will be a critical determinant of success. These sectors, often shackled by stringent compliance requirements and skills shortages, require more than technological promises, they demand proven outcomes. The partnership must therefore navigate a labyrinth of challenges, including bridging expertise gaps and addressing regulatory nuances, to cement its credibility. All while the mainframe platform continues to evolve and modernize to handle transactional AI uses cases with innovative technology such as the Telum II processor and Spyre accelerator. The optionality in the mainframe space is unprecedented and clients have options. Competition and different alternatives aregood for the mainframe ecosystem as a whole.
Going forward I will be focused on tangible evidence of enhanced agility, cost efficiencies, and client satisfaction and ultimately customer examples of full-scope decommissioning before I judge success. Ultimately, this collaboration stands at a crossroads of ambition and execution, with any documented successes being the barometer. What is without question is the broader narrative of mainframe modernization in an AI-driven era.
Steven Dickens | CEO HyperFRAME Research
Regarded as a luminary at the intersection of technology and business transformation, Steven Dickens is the CEO and Principal Analyst at HyperFRAME Research.
Ranked consistently among the Top 10 Analysts by AR Insights and a contributor to Forbes, Steven's expert perspectives are sought after by tier one media outlets such as The Wall Street Journal and CNBC, and he is a regular on TV networks including the Schwab Network and Bloomberg.