IBM z17: Bringing AI to the Core of Business Infrastructure
IBM has introduced its latest mainframe system that strives to ensure businesses can enhance innovation and unlock value across more than 250 AI use cases - IBM z17.
This AI-ready infrastructure is crafted to embed AI into enterprise infrastructure, empowering organisations to run real-time, high-volume AI workloads efficiently and securely.
IBM z17 is powered by the IBM Telum II processor and ensures AI becomes a secure and scalable area of enterprise IT. Compared to z16, IBM z17 offers organisations the opportunity to process 50% more AI inference operations every day.
This mainframe system will transform how businesses make decisions, interact with data and respond to complex operational challenges.
IBM z17 will be generally available from June 2025.
Ross Mauri, GM of IBM Z and LinuxONE, IBM, says: “The industry is quickly learning that AI will only be as valuable as the infrastructure it runs on.
“With z17, we're bringing AI to the core of the enterprise with the software, processing power, and storage to make AI operational quickly. Additionally, organisations can put their vast, untapped stores of enterprise data to work with AI in a secured, cost-effective way.”
How IBM z17 allows enterprises to have AI at the core
IBM z17 allows enterprises to embed intelligence directly into their mission-critical operations by ensuring AI remains at its core.
It acts as a foundational platform for operationalising AI at scale by expressing the AI-first architecture across software, hardware and security.
The Telum II processor, which lies at the heart of IBM z17, features a second-generation on-chip AI accelerator. This can deliver 1 millisecond response times for real-time decision-making and handle more than 450 billion inference operations per day, allowing organisations to run AI models directly where data is processed.
IBM z17 supports real-time inferencing during transactions, allowing businesses to detect fraud within milliseconds, make AI decisions at the point of data generation and score all customer interactions instantly.
z17 utilises tools such as the watsonx Assistant for Z, watsonx Code Assistant for Z and Z Operations Unite to integrate AI across the developer and operator experience. These tools increase operational efficiency and reduce time-to-resolution through AI-driven automation.
The new z/OS 3.2 (which is expected in Q3 2025) supports hardware-accelerated AI insights for system management through AI-based anomaly detection and predictive analytics for system performance.
AI also boosts IBM z17’s security features with features such as:
- IBM Vault
From HashiCorp, this offers identity-based secrets management across hybrid cloud
- IBM Threat Detection for z/OS
This uses AI to identify malicious anomalies
- AI-based data classification
This AI-driven feature helps organisations to find and secure sensitive data
The role of IBM’s Spyre accelerator
The IBM Spyre Accelerator is designed to extend and improve the system’s AI capabilities, especially for Gen AI workloads. This AI compute add-on for the IBM z17 mainframe is expected to be available in Q4 2025 as a PCle card.
The Spyre Accelerator will work with the Telum II processor to support AI assistants and agents, run LLMs and enable Gen AI applications to operate natively on the mainframe.
It maintains control over sensitive business logic and enterprise datasets and aligns with security and compliance requirements for regulated industries by keeping AI workloads on-platform.
By seamlessly integrating with the IBM AI Ecosystem, the Spyre Accelerator will be compatible with AI workloads already being developed for z17 and will execute watsonx-powered AI agents and assistants.
The IBM Spyre Accelerator demonstrates IBM’s drive for making IBM z17 a core hub for enterprise AI by allowing organisations to accelerate time to value for AI-driven services and modernise applications without altering their infrastructure.
- Real-time AI inference at scale
- Enhanced AI compute power
- AI-driven operations and automation
- Generative AI integration
- Unified observability and system insight
Reimagining the role of mainframes in the AI era
IBM z17 will demonstrate how mainframes can now act as intelligent platforms for automation and real-time insight in the dynamic AI era.
By possessing the ability to handle more than 450 billion inferencing operations a day and allowing businesses to score all transactions in real-time, IBM z17 will make AI operational within core workloads.
The Spyre Accelerator will open up new AI-led experiences, including predictive maintenance, customer service bots and advanced document processing.
IBM z17 allows IT teams to work smarter, faster and will have fewer manual interventions by integrating AI into the user experience. Developers can write code and solve incidents using live system data through tools such as watsonx Assistant for Z and IBM watsonx Code Assistant for Z.
Furthermore, AI-powered chat interfaces can also detect and solve issues across the mainframe environment for the first time as watsonx Assistant is now integrated with Z Operations Unite.
IBM z17 redefines the role of the mainframe as a high-performance AI platform that can offer consistent security, speed and scalability.
Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE.
Discover all our upcoming events and secure your tickets today.
AI Magazine is a BizClik brand
