At Hot Chips 2026, IBM announced a collaborative project with Arm Holdings to build a dual-architecture processor capable of executing IBM Z mainframe and Arm-native Linux workloads on a single core, a first in the industry. This leap aims to merge robustness of enterprise systems with the expansive Arm AI software ecosystem.
- First chip to natively run both IBM Z and Arm workloads on same core
- Built on 2nm technology with 11 cores at 5.7 GHz
- Includes AI accelerators for fraud detection, scalable to hundreds of cores
What happened
IBM announced a pioneering partnership with Arm Holdings to develop a dual-architecture processor designed to power its next-generation IBM Z mainframes and LinuxONE systems. This new chip is unique because it can simultaneously run z/OS and Arm-native Linux workloads on the same core, eliminating the need for virtualization or emulation which traditionally hinder performance.
The chip utilizes an advanced two-nanometer process with 11 high-performance cores capable of running at speeds up to 5.7 GHz. It also integrates a dedicated data processing unit for efficient input/output operations and AI inference accelerators tailored for real-time fraud detection in financial transactions. Additionally, the design supports scaling to hundreds of cores to accommodate enterprise growth.
Why it matters
This innovation bridges two historically separate computing ecosystems: IBM’s trusted mainframe infrastructure, which processes approximately 70% of all global financial transactions, and Arm’s expansive software ecosystem that supports cloud-native and AI development with over 22 million developers worldwide. Financial firms and other regulated industries gain a new avenue to run AI workloads securely on-premises while maintaining the reliability and security of IBM mainframes.
By natively supporting Arm workloads without performance-robbing virtualization, IBM’s new chip enables enterprises to leverage cutting-edge AI and cloud-native applications directly on their legacy infrastructure. This addresses long-standing challenges faced by organizations needing highly secure yet flexible platforms for sensitive workloads, making IBM’s systems more adaptable to modern AI-centric business demands.
What to watch next
Market observers will be looking for IBM to reveal the chip’s official name, detailed specifications, and availability timeline. IBM and Arm’s co-designed architecture may set a precedent for future enterprise hardware designs that blend efficiency, security, and AI capabilities. The performance of the chip in real-world workloads, especially its AI inference for fraud detection and scalability, will be key indicators of its impact.
Additionally, the adoption rate among financial services and other highly regulated sectors will be critical to watch. The success of this chip could encourage broader AI deployment directly on premise, potentially reshaping how sensitive data-driven industries approach artificial intelligence and hybrid workload management.