System Z reinvented: Why IBM mainframes dominate enterprise IT

Mainframe computing is having a moment — again. Nearly 60 years after the first IBM System/360 debuted as a mainframe platform, it continues to evolve in ways that defy expectations, powering critical workloads across enterprises. In an era shaped by hybrid cloud, artificial intelligence acceleration and rising regulatory pressure, IBM System Z remains a constant at the center of global business infrastructure.

The staying power of System Z isn’t a coincidence. Seventy-five percent of more than 2,500 global IT executives surveyed by the IBM Institute for Business Value rate mainframes equal to or better than cloud computing in total cost of ownership. Enterprise IT leaders aren’t clinging to nostalgia; they’re making calculated decisions about resilience, security and performance at scale. As organizations wrestle with spiky AI inference demands, ever-tighter data controls and the need for real-time decisioning, the core values of mainframe architecture are becoming increasingly relevant.

This spotlight puts IBM in a unique position. The company’s latest z17 System aims to bridge transactional processing with AI-enabled automation, embedding inferencing closer to the data and fortifying security down to the chip level. It’s a continuation of a modernization arc that’s seen System Z integrate more tightly with cloud-native tools and platform engineering strategies across sectors.

“IBM Z’s advantage, the underlying platform, is an integrated system; IBM’s deep tech … the silicon expertise, the fact that they design their own silicon and they’ve got AI chips on there,” said theCUBE Research’s Dave Vellante during the “IBM Z: Hybrid Cloud With Advanced AI Where It Matters Most” event. “They’ve got Peripheral Component Interconnect Express accelerators … and a data stack through lots of data governance … the security piece. Obviously, IBM mainframe security is best-in-class. These systems don’t go down.”

This feature is part of SiliconANGLE Media’s exploration of the mainframe’s evolving role in modern computing environments. (* Disclosure below.)

Why staying on System Z makes sense

Enterprises have significantly ramped up their digital transformation efforts in recent years, resulting in increased transaction volumes and intensified mainframe workloads. Contrary to outdated perceptions, today’s mainframe technology, such as IBM’s z17, delivers state-of-the-art processing capabilities that surpass alternative systems, according to Ross Mauri, general manager, IBM Z, at IBM. This evolution has made the mainframe indispensable for handling critical, high-value applications, especially in heavily regulated sectors.

“Clients are looking at doing more and more online in real time,” Mauri told theCUBE. “To the consumer, they’re just doing things on their phone. But in the systems that support all those applications, those are all pretty much on mainframes, [which means] more transactions to the mainframe. There have also been regulatory increases over time that banks have to have more resiliency around the world. So, we’re setting up more resilient infrastructures with our clients … the important thing for me now is to continue to introduce new capabilities into that stack so that clients get more and more value out of their mainframe.”

Quantum computing poses a future threat that could potentially crack today’s encryption methods, thereby placing sensitive enterprise data at risk. As mainframe workloads grow alongside digital transformation, protecting this data becomes increasingly urgent. To mitigate this looming threat, Mauri recommends that organizations, particularly those in essential industries, swiftly transition to quantum-safe cryptography. The IBM z17 supports the new quantum-safe cryptographic algorithms recently standardized by the National Institute of Standards and Technology, offering clients proactive protection against future quantum-enabled threats.

“It’s a Y2K moment without that exact moment in time when it’s going to happen, because no one really knows when quantum computers will be large enough from a qubit scale point of view and stable enough to crack 256-byte crypto keys,” Mauri told theCUBE. “We’re just going to call that out there … that when [hackers] break in and they’re stealing encrypted data, they’re just stealing the data and … putting it aside and they’re waiting until they have a quantum computer that can break it. The sooner clients can encrypt with the new quantum-safe algorithms, the more protected they’re going to be.”

The urgency of quantum-safe cryptography illustrates just one dimension of the strategic importance of System Z. Beyond security, analysts highlight the mainframe’s growing role in managing increased transactional demands and diverse workloads, making it not merely defensive, but proactively essential.

The IBM mainframe platform remains deeply embedded in critical industries, according to theCUBE Research’s John Furrier. Its suitability extends beyond traditional sectors, showing significant growth into broader transactional processing roles. With expanding use cases and increasing regulatory complexities, the System Z has a robust architecture that offers unmatched coherence, resilience and transactional efficiency critical for high-stakes environments.

“Seventy percent of all financial transactions go through Z, which is a huge stat,” Furrier said. “People might not know that, but that’s just in financial services. One of the things that’s coming out of this is [that it’s] clear that Z is not just surviving, it’s thriving. Clients are modernizing around Z … a two-times increase in use cases, which shows that Z is starting to venture into … mainstream transaction processing beyond their code base … I think there’s a whole other use case.”

Putting AI at the core: How z17 changes the game

IBM’s z17 mainframe is positioned uniquely in an AI infrastructure landscape largely dominated by graphics processing unit market leaders, such as Nvidia Corp., and major cloud providers, including Amazon Web Services, Inc., Microsoft Azure and Google Cloud, according to Steve McDowell, chief analyst and chief executive officer of Nand Research LLC. The z17 distinguishes itself by providing an on-premises, vertically integrated solution optimized explicitly for enterprises demanding precise data control, minimal latency and heightened transactional efficiency, he noted.

“With the z17, IBM offers an on-premises, vertically integrated solution tailored for industries that demand regulatory compliance, data control and uninterrupted operations,” McDowell wrote in a recent Forbes article. “This strategy gives IBM a unique position in the evolving AI infrastructure landscape.”

Central to the performance leap of the IBM System z17 is the innovative Tellum II microprocessor and Spyre AI accelerators, according to Mauri. Tellum II introduces enhanced capabilities within its on-chip AI acceleration unit, markedly expanding inferencing power. Whereas the previous generation linked each processing core to a single accelerator, the z17 enables each core to simultaneously access eight accelerators, significantly amplifying AI processing throughput. Spyre complements Tellum II by integrating encoder and decoder models tailored for transactional operations and agent-based AI tasks.

“We’ve significantly enhanced the capabilities and performance within this accelerator,” Mauri told theCUBE. “By putting the AI accelerator in the microprocessor, we’ve basically reduced latency as much as you possibly can. Clients can now do multiple AI inferences, each completed in less than a millisecond, within a single transaction. That’s the game-changer.”

The System z17’s AI innovations extend beyond hardware enhancements, integrating IBM’s watsonx platform throughout its software stack to enable multimodal, real-time inferencing for mission-critical applications across diverse industries, according to Tina Tarquinio, vice president of product management design, IBM Z and LinuxOne, at IBM. Embedding accelerators directly on-chip significantly reduces latency, crucial for real-time applications such as fraud detection and anti-money laundering scoring.

“This is a full platform launch — AI technology on the chip, attached Spyre accelerator and AI all the way up through the full software stack,” Tarquinio told theCUBE. “That’s part of our secret sauce: Teams that work on the silicon all the way through the application layer.”

IBM’s new z17 platform stands apart in the era of agentic AI because it’s purpose-built for the next generation of cross-enterprise reasoning and decision intelligence. Where most enterprise AI runs into bottlenecks of scale, trust and compute intensity, the z17 provides unmatched capacity for real-time, high-volume inference, causal reasoning and semantic processing — the foundations of tomorrow’s AI-driven enterprise, according to Scott Hebner, principal analyst for AI at theCUBE Research.

IBM’s long heritage with mainframe architecture has effectively future-proofed its clients for this next era of agentic AI, according to Hebner.

“The mainframe has always been about dependable computation at scale,” he said. “What’s different now is how that foundation enables enterprises to build intelligent agents that can reason, decide and act across systems — safely and in real time.”

The z17’s integration of causal AI, enterprise semantic layers and connected knowledge graphs enables organizations to move beyond predictive analytics into prescriptive, context-aware decisioning. This feature makes z17 uniquely positioned not just as a transaction engine, but as an AI control plane for enterprise knowledge, powering intelligent agents that must reason across industries, regulations and global business ecosystems, Hebner added. In short, the IBM z17 offers enterprises an AI-future-proof investment, providing them with the computational backbone to unlock agentic AI at scale and with trust.

“As enterprises move toward semantic layers, knowledge graphs and causal inference as the backbone of digital decision-making, the IBM z17 stands out as a future-proof investment,” he said. “Its ability to process the most compute-intensive reasoning workloads will be essential for scaling agentic AI across industries.”

Together with IBM’s broader AI strategy under the Magenta brand, the z17 forms a critical pillar of an enterprise’s AI readiness — blending decades of mainframe reliability with modern decision intelligence. It’s a system built not just to run workloads, but to reason through them — setting the stage for the next wave of AI-native, self-optimizing enterprises.

Breaking down barriers: z17’s hybrid and AI integration

IBM’s software strategy for z17 emphasizes seamless integration with hybrid cloud environments and advanced AI-driven operations, according to Skyla Loomis, general manager, IBM Z software, at IBM. Loomis’ team prioritizes reducing complexity and inviting new generations of developers and operators into the Z ecosystem, moving beyond traditional mainframe approaches. IBM is phasing out platform-specific scripting, favoring enterprise-standard tools such as Ansible and Terraform to empower users without deep Z operating system knowledge.

“We want to seamlessly integrate into the hybrid cloud,” Loomis told theCUBE. “We want to empower and embed AI around all of that to dramatically simplify and reduce the skills that are needed to be successful.”

Complementing this modernization, IBM incorporates standard development tools, including Git, Jenkins and Visual Studio Code, directly into the System Z operating system, coupled with AI-powered visibility into application dependencies and behaviors, according to Loomis. The goal is to mitigate risks associated with changes to legacy applications, allowing developers to work confidently within familiar environments.

“I often think of these applications as big tangled-up balls of string, and you’re afraid to pull on one of the strings because you don’t know what it’s connected to,” Loomis explained. “This is where I think AI becomes a game changer for us because we’re able to give you X-ray vision into that ball of string.”

Tarquinio underscores IBM’s deep commitment to customer-driven product evolution, highlighting extensive user research as foundational to the z17’s design and features. With a clear goal of simplifying mainframe experiences and accommodating contemporary skills, the company leverages IBM Design Thinking to prioritize innovations.

“We’ve spent over 2,000 hours with our clients doing user research to come up with what we are going to put in the next-generation mainframe,” Tarquinio said. “We’re really focusing on what they need, when they need it and how they need it.”

IBM’s strategic use of AI and hybrid cloud capabilities aligns with industry trends highlighted by IDC, emphasizing how AI-driven mainframe modernization accelerates digital transformation while reducing associated risks.

“Adopting AI-driven mainframe modernization allows organizations to automate, optimize and expedite their digital transformation, transforming legacy mainframe systems into innovation catalysts while mitigating associated risks,” said Cyrille Chausson, IDC Research Manager, European Application Modernization Strategies, at IDC Research Inc.

The overarching ambition of IBM’s software strategy on z17 is to transform IT operations through comprehensive AI-driven observability, anomaly detection and autonomous remediation. By using agentic AI, IBM aims to enhance transparency and facilitate operational understanding, not merely automating tasks but enabling greater expertise among IT professionals.

“With agentic AI, they can talk to each other,” Loomis concluded. “We want to get to also providing the explanation of everything that happened. So, there’s transparency — we’re not just helping them catch their fish. We’re teaching them how to fish.”

(* Disclosure: TheCUBE is a paid media partner for the “IBM Z: Hybrid Cloud With Advanced AI Where It Matters Most” event. Neither IBM Corp., the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Image: SiliconANGLE/Bing