Today, IBM unveiled the IBM z17®, a next-generation mainframe powered by a robust AI infrastructure. By harnessing AI, IBM is transforming the mainframe experience, driving productivity gains for developers, operators and business users.

In a recent survey, 88% of global IT executives stated that application modernization is a crucial step in their journey, and 78% affirmed that mainframes will remain central to digital transformation. Aligned with our clients’ needs, IBM is delivering new innovations to drive even more value from their transactional processing applications through AI-driven productivity and efficiency, streamlining on-boarding of new talent, and empowering clients to achieve more for their business.

The IBM z17 and Z software provide a transaction processing platform that delivers unparalleled performance, security and reliability, infused with AI-driven insights. This integrated platform delivers a comprehensive set of AI capabilities that includes AI assistants and AI agents supporting Application Development, Application and Data Management, and Operations Management.

We’re excited to highlight some of our new software announcements for 2Q 2025 in support of z17.

Transaction processing applications are at the core of our client’s business. Our objective is to help them accelerate mainframe application modernization using generative AI and DevOps automation, boosting developer productivity and business agility. We have over 250 completed or ongoing generative AI projects with clients to help them unlock greater productivity and efficiency with IBM Z.

• With application development, everything starts with the developer. Providing the right AI-enhanced developer experience is the priority. In late 2023, we delivered IBM watsonx^TM Code Assistant for Z to accelerate mainframe application modernization using generative AI. And now, integrated AI chat and agentic AI capability allows developers to quickly pinpoint where business logic resides to simplify and streamline application enhancements. We are further extending our use of generative AI for code generation and explanation, and automated refactoring to additional programming languages. As businesses increase application coding agility, AI-assisted test automation further accelerates the modernization lifecycle. IBM Test Accelerator for Z streamlines mainframe application testing, reducing costs and improving quality. For functional testing, Test Accelerator for Z will integrate with watsonx^TM Assistant for Z to enable automatic test generation using natural language prompts. In addition, Test Accelerator for Z will infuse AI into unit and integration testing.
• Mainframe applications have been developed over decades. To fully exploit the innovations of new mainframe technology like z17 requires software and compilers to be at current versions. For example, getting to the latest version of COBOL is necessary to exploit all the AI innovations available.  To accelerate that journey, IBM has introduced the COBOL Upgrade Advisor for z/OS to help businesses upgrade to IBM Enterprise COBOL for z/OS 6 with automated analysis and reporting using a modern VS Code interface (generally available 9 May 2025).

Overall, these announcements aim to boost developer productivity, simplify mainframe application development and drive business agility leveraging AI.

At its core, the mainframe is a secure transaction processing environment and the definitive source of an organization’s most current and valuable data. Ensuring this data is readily accessible to modern hybrid cloud and AI applications is crucial to help our clients drive new business opportunities and fuel growth.

Below is a highlight of several new enhancements designed to help clients extract even greater value from their mainframe data for AI insights:

• Unstructured data is an estimated 90%¹ of all enterprise data making it a critical differentiator for delivering greater AI value. However, tapping into this rapidly expanding data source and combining it with structured data can be challenging. New capabilities for Db2 for z/OS users will provide deeper and more intelligent AI insights from both structured and unstructured data. The SQL Data Insights capability in Db2 for z/OS enables deeper AI insights from structured enterprise data without the need for data science skills. SQL Data Insights has been enhanced with new REST APIs and shell CLIs to simplify automation of model training and key operations, harnessing the IBM Telum II on-chip AI accelerator for faster, smarter decision making. In addition, IBM intends to deliver advanced vector database capabilities for Db2 for z/OS that will exploit the z17 integrated AI accelerator unlocking the next wave of insights by combining structured and unstructured data for more accurate and comprehensive AI insights.
• Further harness the power of transactional mainframe data for AI. We are introducing a new bundle of IBM watsonx.data with IBM Data Gate to make it easier for clients to synchronize Db2 for z/OS, VSAM and IMS transactional data with a watsonx.data lakehouse for enhanced analytic and AI insights. And to boost analytic query performance from mainframe data – IBM intends to deliver a new version of Db2 Analytics Accelerator for z/OS (Accelerator) that will leverage z17’s 40% cache growth on chip to increase query performance. As part of a phased delivery plan, Large Object (LOB) data will be accelerated in the same way as traditional structured data. Accelerator to Accelerator data copying will be possible enabling tables to be copied between Accelerators to support high availability and workload balancing across accelerators.

Providing a security-rich environment for IBM Z data and applications is part of the mainframe’s DNA. The new capabilities from HashiCorp help standardize secrets management across hybrid cloud, are now available on IBM Z. In addition, IBM Vault Self-Managed for Z and LinuxONE provides identity-based security to automatically authenticate and authorize access to secrets. With the addition of IBM Vault, clients now have a single solution for protecting critical workloads by securely managing secrets across their full enterprise technology estate.

Because of the business criticality of mainframe applications, maintaining service levels and business resiliency is paramount. Our goal is to transform and streamline operations with AI-enhanced anomaly detection, application observability and automation leveraging Open Telemetry as part of a comprehensive resilience management approach. Let me share a few of our recent innovations in operations management:

• New IBM Z Operations Unite is a transformative solution that unifies IBM Z operational data from multiple sources, streamlines operations with AI and reduces alert investigation by 9X². By accelerating anomaly detection, isolating incidents, and cutting resolution times, IBM Z operations teams can better achieve service level objectives. IBM Z Operations unite will be generally available May 30, 2025.
• To gain deep-dive health and performance insights into IBM z17 we have enhanced IBM Z® IntelliMagic Vision for z/OS®. This will help clients protect application availability and optimize performance with out-of-the box, interactive insights into IBM Spyre® Accelerator usage, I/O optimization, and power consumption. It will also ensure optimal workload performance during and after upgrading to z17 with expert guidance.
• Maintaining operations and business resiliency extends beyond IBM Z to the entire hybrid cloud environment. Using AI, IBM Concert uncovers crucial insights about your entire enterprise operations and provides application-specific recommendations for improvement including IBM Z. Our support for IBM Z includes simplified software maintenance. By automating APAR tracking, risk assessment, and remediation, organizations can prioritize fixes, reduce manual efforts, and improve efficiency and reliability. Available now, IBM Concert for Z helps ensure a proactive and secure IBM Z environment.

• Simple, quick Generative AI adoption: Announcing IBM AI Optimizer for Z empowering organizations to adopt and scale Gen AI workloads for the mainframe with speed and simplicity. Initially, the AI Optimizer for Z will simplify, automate, and accelerate the installation and configuration of IBM watsonx Code Assistant for Z and IBM watsonx Assistant for Z.
• AI Assistants on IBM z17: Coming in 4Q 2025, IBM Spyre Accelerator is expected to enable generative AI on z17. IBM watsonx Code Assistant for Z and watsonx Assistant for Z will leverage the Spyre Accelerator to run natively on IBM z17, offering clients the ability to accelerate application lifecycles and simplified mainframe operations with high levels of security and quality service. By running generative AI on the mainframe with IBM Spyre Accelerator, we simplify architecture, reduce latency, and eliminate external dependencies. Plus, with agentic AI, we offer more secure, end-to-end processing, keeping sensitive data and requests within the mainframe’s secure environment. Availability and procurement of GPUs is at best difficult, resulting in many having to resort to the cloud to rent them or a long lead time in procurement. Spyre Accelerator will make it very accessible — offering faster time to value as there is no wait to access and unleash the power of Gen AI. IBM intends to run IBM watsonx Code Assistant for Z and watsonx Assistant for Z on-premises on mainframe with the availability of Spyre Accelerator in 4Q 2025.

AI is fundamentally transforming and simplifying the ways users experience the mainframe to reduce complexity and help alleviate skills concerns. IBM Z Software is exploiting z17’s AI differentiation to accelerate value capture. And finally, IBM is using AI today to deliver business value around efficiency, productivity and resiliency.

The best is yet to come.