IBM z17 Sets a New Standard for Enterprise AI with Advanced Hardware, Software, and User Experience

IBM has unveiled the IBM z17, the next evolution of its flagship mainframe platform, purpose-built to power enterprise-scale AI applications. Featuring deep integration of AI capabilities across its architecture, the z17 is driven by the new IBM Telum® II processor and introduces powerful tools for generative AI, large language models (LLMs), and intelligent automation.

Next-Level AI Processing with IBM Telum® II

The z17 marks a leap in performance, delivering up to 50% more AI inference operations daily compared to its predecessor, the z16. With over 250 supported AI use cases—including loan risk assessment, chatbot optimization, medical image analysis, and retail fraud prevention—the z17 positions businesses to analyze and act on 100% of transactions in real time.

Central to this performance is the second-generation on-chip AI accelerator within the Telum® II processor. With increased frequency, 40% more cache, and the ability to perform over 450 billion inference operations per day with 1ms response times, this architecture makes real-time AI a practical reality.

Expanding AI Capabilities with IBM Spyre™ Accelerator

Coming in Q4 2025, the IBM Spyre™ Accelerator will add further AI compute capacity via PCIe card. This innovation supports multi-model AI environments and brings generative AI directly to the mainframe—ideal for running AI assistants and agents using native enterprise data.

AI-Powered User Experience

IBM z17 introduces watsonx Code Assistant for Z and watsonx Assistant for Z, enhancing developer and IT operations with AI-generated recommendations and chat-based system management. These assistants integrate directly with IBM Z Operations Unite to detect and resolve incidents using real-time systems data.

Fully Integrated System for Hybrid AI Workloads

IBM z17 is designed for hybrid cloud environments, blending hardware innovation with AI software tools and support for open standards. This foundation allows seamless integration with modern workloads and offers key upgrades including:

  • z/OS 3.2: Launching in Q3 2025, the next version of IBM’s operating system adds support for AI hardware acceleration, NoSQL databases, hybrid data access, and operational insights to improve AI workflows.
  • IBM Z Operations Unite: Available May 2025, this platform collects performance metrics using OpenTelemetry to streamline IT operations and speed incident response. When combined with IBM Concert, it offers enterprise-wide AI-driven correlation of operational data.
  • Granite Model Integration: Enables AI assistants and agents to run securely on-platform using IBM’s proprietary models without external data transfers.

Strengthening Security with AI and Integrated Tools

IBM z17 builds on a legacy of enterprise-grade resiliency with new security features:

  • IBM Vault: Based on HashiCorp Vault, this solution provides identity-based secrets management for the entire hybrid cloud landscape.
  • AI-Powered Data Discovery: Using natural language processing and Telum® II, z17 can discover and classify sensitive data for improved governance.
  • IBM Threat Detection for z/OS: Leverages AI to identify potential cyber threats through anomaly detection.

Advanced Support and Agile Storage for Mission-Critical Data

IBM’s Technology Lifecycle Services offer AI-enabled support for IBM z17, streamlining incident remediation and accelerating case resolution. On the storage front, the new IBM Storage DS8000 (10th Generation) complements z17 by delivering secure, high-performance access to data with a flexible, modular architecture to accommodate evolving business needs.

Availability

The IBM z17 will be generally available on June 18, 2025, with the Spyre™ Accelerator launching in Q4 2025. For full product details, visit the IBM z17 product page.

Explore AITechPark for the latest news and insights on AI, IoT, cybersecurity, and emerging technologies.