IBM z17: Mainframe Meets Artificial Intelligence
AI needs secure and efficient computing power - IBM delivers it with the z17, the first mainframe of a new era. Introduced on April 8, 2025, the IBM z17 is optimized for artificial intelligence. It processes sensitive data directly and securely on site, without the cloud - and saves energy with maximum performance thanks to specialized AI accelerators.
Mainframes have been the backbone of corporate IT for decades, especially in industries such as financial services and insurance as well as in the public sector. They are characterized by their unsurpassed reliability, security and scalability[i]. The IBM z17 is revolutionary because it integrates AI directly into the core architecture[ii]. Compared to its predecessor, the z16, it enables the processing of AI workloads directly at the point of transaction. The IBM z17 marks a turning point in the evolution of mainframe technology. With AI integration, optimized usability, strong security features and high energy efficiency, it sets new standards for business-critical workloads.
The evolution of the mainframe: Why the z17 is unique
The core of the z17 is the new IBM Telum II processor, which is equipped with an integrated AI accelerator. This makes it possible to perform up to 450 billion inference operations per day with a response time of just one millisecond (inference operations refer to the execution of a trained AI model on a processor in order to make predictions or decisions based on new data). This corresponds to a 50 percent increase in performance compared to the z16[iii]. The Telum II is complemented by the optional IBM Spyre™ accelerator card, which will be available from the fourth quarter of 2025 and is specially optimized for generative AI and large language models (LLMs). This combination of Telum II and Spyre enables up to a 7.5x increase in AI performance over the z16, making the platform the ideal system for organizations looking to deploy AI at scale.
AI in action: More than 250 use cases
The IBM z17 was developed to support a wide range of AI applications - IBM speaks of over 250 specific use cases[iv]. The z17 enables organizations to deploy AI not as an isolated technology, but as an integral part of their business-critical processes. For example, in credit underwriting, the system can analyze credit risk in real time by embedding AI models directly into transaction processes. By combining machine learning with large language models, the Z17 enables highly accurate fraud detection in real time, which is particularly beneficial for banks and retailers.
AI-powered assistants, such as the IBM watsonx Assistant for Z, can help businesses manage chatbot services more efficiently and improve the customer experience. In the medical field, the system can use AI models to analyze medical images to aid in diagnosis.
The z17 makes it easier to integrate AI assistants
The z17 supports IT teams and developers with built-in AI assistants. The watsonx Code Assistant for Z streamlines coding, while the watsonx Assistant for Z works with Z Operations Unite to enable chat-based problem detection. Both tools use live data to quickly resolve issues. With many mainframe professionals expected to retire by 2035, IBM is using AI-powered tools on the z17 to help new professionals become proficient more quickly. The z17 is designed to meet the needs of a changing workforce.
Security, energy efficiency and strategy
Security is a hallmark of IBM Z systems, and the z17 is no exception. By processing AI workloads directly on the chip or within the Spyre accelerator, sensitive data remains within the secure mainframe environment, minimizing the risk of data leakage. This is particularly important in industries where data protection and compliance are paramount.
Another advantage is energy efficiency. This is a decisive factor at a time when sustainability and cost efficiency are becoming increasingly important for companies.
The z17 is strategically important because it enables the use of AI on a large scale - seamlessly integrated into existing IT landscapes. This protects investments and accelerates innovation. Instead of transferring data to the cloud, the system enables AI workloads to be processed directly on site. This reduces latency times and increases data security, making the platform particularly attractive for industries with high security requirements, such as banking, insurance, and public administration[v].
The z17 and the future of corporate IT
The IBM z17 underscores the continued relevance of mainframes in an AI-driven world. With its ability to handle generative and predictive AI workloads, support multi-model applications while ensuring the highest security standards, the mainframe system positions itself as a platform for organizations looking to combine AI, innovation and stability.
For organizations looking to drive their digital transformation, the z17 offers a unique opportunity to integrate AI into their core processes without compromising on security or efficiency. At a time when the demands on IT infrastructures are constantly increasing, the z17 is a clear statement from IBM: Mainframes are not only part of the past, but also of the future of corporate IT.
UMB AG: Incomparable mainframe know-how
The IBM z17 will be generally available from June 18, 2025, while the Spyre accelerator cards are scheduled to start shipping in the fourth quarter of 2025[vi]. In addition to hardware, IBM is also introducing new software solutions, including IBM Z Operations Unite in May 2025 and the new version of the mainframe operating system z/OS 3.2 (in the third quarter of 2025). These components form a fully integrated technology stack for maximum flexibility and performance.
UMB is the first and largest IBM Platinum Business Partner in Switzerland with comprehensive expertise in mainframe technology. The IBM mainframe is still regarded as an unsurpassed reliable and powerful workhorse. We offer services including modernization, support and cloud solutions. As the largest IBM partner in Switzerland, UMB has built up unparalleled mainframe expertise over the last 30 years. You can count on our specialists. Get in touch with us.
[i]IBM Mainframe Life Cycle History
[ii]AI Meets The Mainframe: Inside IBM’s Bold Z17 Bet
[iii]IBM boosts mainframes with 50% more AI performance
[iv]Transforming and simplifying the mainframe for greater productivity and efficiency with AI
