IBM recently announced that its new IBM z17 hardware and the updated z/OS 3.2 operating system will be generally available in 2025, with some features scheduled for later release. The z17 systems will offer integrated and enhanced AI capabilities across its hardware, the operating system, and system operations.
Designed for AI processing, z17 systems are engineered to drive real-time AI at scale, enable 100% transaction scoring, and unlock new levels of operational intelligence.
Below is a summary of the new AI capabilities IBM has officially announced for the z17, along with the release schedule for each feature.
Related reading: IBM z17: The First Mainframe Fully Engineered for the AI Age (IBM)
Enterprise AI at Scale: Telum II Processors & Spyre Accelerators
Per IBM, z17 systems can support over 250 AI use cases across many industries, including risk management, fraud detection, anti-money laundering, cybersecurity, and compliance.
The z17 adds these enhanced hardware technologies to provide real-time, on-platform, and scalable AI processing.
Telum II Processor (Available June 2025): The Telum II is the next-generation z system processor. It builds on and expands the AI capabilities of the z16 Telum processor set. Telum II offers more advanced AI features, including built-in Data Processing Unit (DPU) capabilities, integrated AI acceleration, eight 5.5 GHz processing cores per chip, and a significantly larger cache memory per core (up to 40% increase versus IBM z16).
The Telum II accelerator runs at 24 Trillion Operations per Second (TOPS), a fourfold (4x) performance improvement over the original Telum processor.
IBM Spyre Accelerator (available 4Q 2025): Spyre Accelerator modules expand AI compute capacity beyond what Telum II alone can deliver. Built on a 5nm process, each Spyre card will contain 32 AI accelerator cores optimized for machine learning and large language models (LLMs). Multiple Spyre cards can be connected via PCIe to the z17 I/O subsystem to scale and increase AI power as needed. Spyre processors also use lower-precision numeric formats (such as int8) to run LLMs with lower power and memory consumption.
With z17 servers, organizations can deploy AI at scale. According to IBM, z17 systems can process 50 percent more AI inference operations per day than with z16 machines. Organizations using Telum II processors combined with Spyre Accelerators can add significant amounts of AI acceleration, throughput, and performance to IBM Z servers.
Related reading: New IBM z AI Hardware and Processing Trends
AI-Enabled z/OS 3.2 Updates
With the z17 announcement, IBM also previewed z/OS 3.2 (available 3Q 2025), the newest version of the IBM Z operating system. z/OS 3.2 was designed to support newer hardware-accelerated AI capabilities as well as to provide AI-enhanced insights for system management. More importantly, version 3.2 provides support for modern data access methods, NoSQL databases, and hybrid cloud data processes.
These new capabilities are designed to “help AI software tap into a broader set of enterprise data and derive predictive business insights,” which should be critical in helping IBM Z AI analyze unstructured data on hybrid cloud platforms, alongside IBM Z databases. These new capabilities should allow businesses to uncover hidden trends, strengthen compliance monitoring, unlock predictive analytics, and provide better access to enterprise data. z/OS 3.2 is expected to run on select IBM z17 and z16 servers. It may also be compatible with certain IBM z15 models. Check with IBM to see whether your server is compatible with 3.2.
Related reading: How AI Can Transform IBM Z Report Management Capabilities
IBM Z AI-Enabled Tools
In conjunction with the z17 announcement, IBM is also promoting and releasing updated AI tools intended to provide (in IBM’s words) a “User Experience Reimagined via new [AI] Assistants and Agents.”
IBM Z Operations Unite will be upgraded to consolidate performance data, logs, and telemetry from multiple IBM Z sources, using OpenTelemetry format. IBM Z Operations Unite is an AI-powered platform that simplifies mainframe operations by unifying workflows, prioritizing events, and assessing business impact. Powered by AI and used with IBM Concert, it allows teams to detect anomalies, isolate the impact of potential incidents, and reduce resolution time. According to IBM, this upgrade will “streamline IBM Z operations with AI.”
For DevOps users, IBM is also integrating watsonz Assistant for Z with IBM Z Operations Unite to provide AI chat-based incident detection and resolution using live systems data. Additionally, watsonx Code Assistant for Z can accelerate mainframe application development and modernization using Generative AI and automation.
IBM Z Operations Unite upgrades are scheduled to be available in May 2025.
A Pivotal Moment for IBM Z Systems
The z17 marks a pivotal moment in the evolution of IBM Z, seamlessly integrating scalable AI capabilities—across hardware, operating systems, and software—into its core operations. IBM Z is redefining its role in the digital landscape, not as a legacy system but as a strategic platform for driving innovation.
For more information on how IBM Z servers can support your organizational needs—including JCL, Java batch processing, and report and output management—please feel free to contact SEA.
