Dive Brief:
- IBM is collaborating with Arm to build dual-architecture hardware that supports IBM Z mainframes and Arm applications, helping enterprises run AI and data-intensive workloads, IBM said in a Thursday announcement. The deal reflects ongoing efforts from IBM to bolster its mainframe capabilities as “AI moves deeper into core business operations,” the companies said.
- The dual-architecture hardware expands mainframe use cases for enterprises, making the platform more attractive to CIOs, according to Alessandro Galimberti, VP analyst at Gartner. The collaboration benefits existing customers by enabling continued use of IBM Z mainframes without making code changes, while new customers with workloads based on Arm will be able to run the workloads on IBM Z mainframes, Galimberti said.
- The hardware will also reshape VMware migrations for CIOs looking at their options, easing existing barriers to migrating workloads to the IBM Z mainframe, according to Galimberti. “If you look in the long-term for the CIO, that could be a good VMware alternative,” he said.
Dive Insight:
IBM’s collaboration with Arm comes as cloud providers funnel investments into infrastructure, including Arm architectures, in a bid for greater compute power at a lower cost to support enterprise AI adoption.
AWS Graviton, Microsoft Azure Cobalt and Google Axion are all Arm-based custom processors aimed at providing competitive cost and energy efficiency for cloud and AI workloads. Even as AI workloads transition from training to inference on large language models, which is expected to get cheaper over the next four years, costs will continue to rise for enterprises as workloads become even more complex and energy-driven.
Arm’s collaboration with IBM builds on the extension of the semiconductor company’s ecosystem into enterprise environments, “giving organizations greater flexibility in how they deploy and scale these workloads,” Mohamed Awad, EVP of the cloud AI business unit at Arm, said in the IBM announcement.
“As enterprises scale AI and modernize their infrastructure, the breadth of the Arm software ecosystem is enabling these workloads to run across a broader range of environments,” Awad said.
The IBM and Arm collaboration is focused on the expansion of virtualization technologies to allow Arm-based software environments to operate in IBM’s enterprise computing platforms, supporting demands of modern workloads and expanding flexibility for how applications in software ecosystems are deployed, according to the announcement.
According to Galimberti, the dual-architecture hardware establishes IBM's mainframe as a viable alternative to VMware, which was acquired by Broadcom in November 2023. Although enterprises haven’t exited VMware en masse, 86% of IT decision-makers said their organizations are “actively reducing their VMware footprint,” according to a recent CloudBolt Software survey.
IBM is investing heavily in bringing AI innovation to the mainframe, including its Telum II processor and Spyre Accelerator, Galimberti said. Now, the dual-architecture hardware offers a “completely different way to see computing overall,” he said.
Businesses are looking for platforms that can evolve alongside the continued development of AI and data-intensive applications “without forcing disruptive tradeoffs,” Patrick Moorhead, founder, CEO and chief analyst at Moor Insights & Strategy, said in the IBM announcement.
“What IBM and Arm are signaling here is a meaningful step toward that future that could broaden how enterprises think about deploying and scaling modern workloads,” he said.
There is not a huge push among CIOs to move off the mainframe, Galimberti said. But CIOs are facing mounting pressure to use the “mainframe in a different way than before” as they’re being asked to implement AI and a modern developer platform.
“I still have some people that try to move off the mainframe and I’m hearing all these horror stories,” he said. “Mainframe exit failures are the best kept secret in IT now.”
