IBM has just rolled out its latest innovation, the z17 mainframe, which is set to revolutionize AI integration across a multitude of applications. Announced on a recent Monday, this cutting-edge system is powered by the IBM Telum II processor, supporting over 250 AI use cases, including AI agents and generative AI. It’s no surprise that mainframes are still a staple for 71% of Fortune 500 companies, underscoring their lasting impact in the tech world.
By 2024, the mainframe market was valued at around $5.3 billion, according to Market Research Future. The z17 is a powerhouse, capable of handling 450 billion inference operations every single day—an impressive 50% improvement over its predecessor, the IBM z16, which launched back in 2022.
One of the standout features of the z17 is its seamless integration with various hardware, software, and open-source tools, addressing the ever-evolving demands of AI. Tina Tarquinio, IBM’s VP of Product Management and Design for IBM Z, shared that development on the z17 kicked off five years ago, well before the AI boom triggered by OpenAI’s ChatGPT in late 2022.
IBM poured over 2,000 hours into research, gathering insights from more than 100 clients to shape the z17’s features. Tarquinio mentioned, “It’s been quite a ride knowing we’re introducing an AI accelerator, and then witnessing all the industry shifts in AI, especially in the latter half of 2022.”
The z17 is built to adapt as the AI landscape continues to change. Initially, it supports 48 IBM Spyre AI accelerator chips, with plans to double that to 96 within a year. Tarquinio emphasized, “We’re building in headroom…ensuring we’ve got the capacity for bigger, larger models.”
Energy efficiency is another key feature of the z17, offering a seven-and-a-half-fold increase in AI acceleration while consuming significantly less power than similar platforms. Tarquinio elaborated, “On-chip, we’re boosting AI acceleration by seven and a half times, but that’s five and a half times less energy.”
