IBM z17 Puts AI at the Heart of Business Technology
IBM’s z17 mainframe positions itself as the next evolution in enterprise IT by placing AI at the centre of infrastructure.
With the IBM Telum II processor and support for the upcoming Spyre Accelerator, IBM is building a platform where business-critical applications can run AI workloads securely and at scale.
The company aims to move beyond traditional data processing and into an era of real-time decision-making, intelligent automation and seamless integration with Gen AI tools.
Embedding AI into mission-critical systems
With IBM z17, AI is no longer an added feature – it becomes an integral part of the infrastructure.
This system is built to serve more than 250 AI use cases across industries, delivering high-volume inference directly within core business transactions.
At the heart of z17 lies the Telum II processor, equipped with a second-generation on-chip AI accelerator.
This setup allows enterprises to run models where their data lives, offering latency as low as one millisecond for real-time decisions.
In practice, this translates to more than 450 billion inference operations processed every day.
Businesses can use these capabilities for instant fraud detection, real-time customer interaction scoring and transactional insight generation.
The on-chip acceleration supports AI inferencing during transactions without interrupting business flows. In doing so, z17 enables enterprises to embed intelligence directly into their workflows – a critical step in operationalising AI across large-scale environments.
IBM z17 also integrates with tools such as watsonx Assistant for Z and watsonx Code Assistant for Z.
These tools assist developers and operators in automating tasks, improving system uptime and speeding up incident response.
The mainframe uses Z Operations Unite to unify monitoring and management, reducing complexity across mainframe environments through AI-based support.
A platform built for AI innovation and resilience
Security, scalability and performance remain pillars of the IBM z17.
New AI-enabled features within z/OS 3.2, expected in Q3 2025, will provide anomaly detection and predictive analytics, enhancing system resilience.
These capabilities allow IT operations to predict system behaviour and prevent issues before they escalate.
The system also offers a suite of AI-based security features:
IBM Vault from HashiCorp delivers secrets management based on identity for hybrid cloud setups
IBM Threat Detection for z/OS employs AI to identify abnormal activity within workloads
AI-based data classification locates and protects sensitive data using machine learning models.
By combining infrastructure-level intelligence with high-performance processing, IBM z17 serves as a central platform for building, deploying and scaling enterprise AI applications.
Ross Mauri, GM of IBM Z and LinuxONE, IBM, emphasises the importance of AI-ready infrastructure: “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.”
- Real-time AI inference at scale
- Enhanced AI compute power
- AI-driven operations and automation
- Generative AI integration
- Unified observability and system insight
The Spyre Accelerator and future-facing enterprise AI
To expand the AI potential of z17, IBM will introduce the Spyre Accelerator in Q4 2025.
This plug-in component, deployed via a PCle card, is designed to handle Gen AI workloads more efficiently and aligns with stringent security and compliance standards.
By working alongside the Telum II processor, the Spyre Accelerator will support LLMs, AI agents and assistants.
It enables these models to run natively on the mainframe without external compute dependencies, which keeps enterprise data secure and ensures compliance within regulated industries.
Compatible with the IBM AI Ecosystem, the Spyre Accelerator allows businesses to accelerate development by running watsonx-powered assistants directly within their infrastructure.
This approach avoids the need to move sensitive data off-platform and supports Gen AI innovation without requiring architectural changes.
IBM is positioning the Spyre Accelerator not just as an add-on but as a strategic element for enterprises aiming to modernise applications and launch AI-driven services faster. It reinforces IBM’s aim to make the z17 a central hub in an AI-native enterprise environment.
Redefining mainframes for the AI-first enterprise
IBM z17 signals a new phase in enterprise computing where mainframes evolve from traditional data processors to intelligent platforms for AI-powered automation and analytics.
With its ability to score transactions in real time and integrate directly into developer and operations workflows, z17 is geared for the demands of the AI era.
Developers can now code using live system data, aided by AI interfaces such as watsonx Assistant.
AI-based chat tools embedded into Z Operations Unite allow for interactive problem-solving across the infrastructure.
IT teams gain the tools to reduce manual intervention, automate processes and act on insights as they emerge.
In doing so, IBM z17 reinvents the role of the mainframe – not just as a legacy system, but as a modern, high-speed, high-security AI platform for innovation.
Explore the latest edition of Technology 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.
Technology Magazine is a BizClik brand
