With the launch of IBM Z16 in April 2022, the mainframe platform began its journey into on-chip AI processing through the Telum processor.
With the introduction of IBM Z17 in April 2025, the platform has significantly scaled its native AI processing capabilities, delivering a 50% increase in performance. The Telum II processor (on-chip), supported by the Spyre Accelerator (PCIe-attached), offers a fourfold increase in compute power, achieving up to 24 trillion operations per second (TOPS). This enables IBM Z17 to process over 450 billion inference operations per day with a response time of just one millisecond.
Enhancing business functions with AI capabilities
IBM Z17, combined with z/OS 3.2, provides advanced capabilities for real-time AI-driven fraud detection. For example, banking transactions can be scanned for suspicious activity during processing, triggering immediate alerts.
Modernization efforts now naturally incorporate AI and Generative AI (GenAI). Several mainframe software vendors have introduced AI-enabled features, including chatbots, runtime log and event analysis, automated fix suggestions and integrated resolution options,all accessible from a single interface.
These enhancements enable the integration of requirement-specific AI assistants and agents using IBM’s Granite model, which is natively supported on IBM Z17 hardware, thereby eliminating the need to move data off-platform and reducing risk.
z/OS 3.2 strengthens cyber-resiliency through quantum-safe cryptographic algorithms and Threat Detection for z/OS (TDz). It also leverages AI-infused tools such as the AI-powered Workload Manager (WLM), which predicts workload spikes and proactively adjusts resources to meet demand.
AI Co-Pilot for technical support teams
Watsonx Assistant for IBM Z empowers support teams to troubleshoot, administer and manage system-level tasks efficiently,saving time and significantly reducing the learning curve for technical staff. It can also trigger Ansible scripts to execute required actions.
While z/OS 3.1 introduced AI-powered workload management for batch initiators, z/OS 3.2 enhances this capability by predicting patterns based on service unit consumption and initiating batch processes more rapidly.
The z/OSMF AI Control Interface provides visualization tools that allow users to compare active initiators with predicted ones in simulation mode. Additionally, z/OS 3.2 includes the SMF Explorer, which provides deeper visibility into system metrics while keeping SMF data within the mainframe, eliminating the need to offload data to external systems for analysis
BMC’s AMI Ops solution further supports observability and automation, offering dashboards to detect anomalies effectively.
Seamless integration with Hybrid Cloud
Organizations are increasingly seeking to leverage the benefits of the cloud while maintaining the reliability and security of their mainframe systems. IBM Z17 with z/OS 3.2 is a cloud native platform that enables containerized workloads via OpenShift and integrates DevOps pipelines through Dependency-Based Build. Infrastructure provisioning is supported through Infrastructure as Code using Ansible and Terraform.
z/OS 3.2 also introduces RESTful APIs via Cloud Data Access (CDA), allowing cloud-native applications to read and write to Cloud Object Storage without requiring ETL processes. This enables modern developers to build applications in the cloud while executing them on the mainframe.
Fostering modern work practices
IBM Z17 and z/OS 3.2 support modern development toolchains, enabling next-generation developers to move beyond traditional green screens, such as ISPF. Developers can now utilize tools such as VS Code, IBM Developer for Z (IDz), GitHub, Jira, Jenkins, Azure DevOps, and more, integrating with or replacing legacy source code management (SCM) systems.
Organizations are increasingly adopting Java for mainframe development. A recent vendor survey revealed that Java is now the most frequently used language on IBM Z, with usage increasing by 49% over the past two years.
Conclusion
The IBM Z ecosystem is already modernized. Organizations must now upgrade and transform their mainframe environments to leverage the latest capabilities in AI, GenAI and hybrid cloud. This transformation may involve in-place modernization or integration with cloud platforms. With modernized mainframes, all of this is achievable. HCLTech offers proven methodologies and expertise to assess client environments and deliver tailored modernization strategies,recognizing that a one-size-fits-all approach does not apply.
