Enterprise IT carries a long list of jobs that nobody notices until something goes wrong; from patching firmware to watching out for the ransomware attack that statistically will come. New IDC research shows 85% of organisations now see generative AI as a major new workload, on the same scale as ERP or e-commerce, yet most are still testing it rather than running it, held back by outdated infrastructure and bottlenecks in CPU speed, memory and networking.
IBM built Power11 to help enterprises stay ahead as AI reshapes how they operate. “Most of the demand for Power11 here is coming from financial institutions and existing Power customers moving up from Power10,” says Greg Christoforides, Technical Director at Anykey Technologies. “Moving to Power11 protects the investment they have already made and lets them stay on the IBM Power platform with continued support.”
Closing the gap
Keeping a Power system current used to mean accepting downtime somewhere along the line, whether through firmware updates, I/O adapter changes or virtualisation software upgrades. Power11 is designed to help reduce that trade-off through live kernel fixes, dynamic firmware updates and automated workload mobility, so critical fixes go in while the system keeps running.
“The reduction in planned downtime in Power11 comes from software that orchestrates the maintenance, checking firmware and checking for security exposures,” explains Christoforides. “It then schedules the work and automates the rollout, moving workloads from one system to the next.”
For financial institutions driving local demand, that matters less as a feature and more as protection for years of investment. “It is also about the continued support of the products they already have and the continuation of the IBM Power platform,” he adds. Done correctly, it is designed to help move much planned maintenance into automated background processes.
A ticking clock
IBM has also rebuilt Power11’s encryption to be quantum safe. Quantum computing, says Christoforides, has moved encrypted data from "classically” safe to "post-quantumly" exposed. “Until quantum computing becomes cryptographically relevant, cracking today’s keys could take years. But when cryptographically relevant quantum computers come online – which could be in just a few years – cracking RSA, ECC and other standard protocols could be reduced significantly,” he explains.
The risk sits in the gap between today and whenever that quantum capability becomes available, because a bad actor with stolen (or “harvested”) data now, even while encrypted, can simply wait until they have access to the quantum technology to decrypt it. “Customers ask me why they should be concerned right now, and the honest answer is that current encryption remains effective today,” he continues. “But quantum computing is quickly advancing towards cryptographic relevance, and organisations should plan now for its potential impact on data protection.”
The same logic sits behind Power11’s rapid ransomware threat detection, which in IBM testing is designed to identify threats in under one minute.
Power11’s AI runs on two tracks, specialised hardware for the heaviest workloads and a development tool built to save time on the most routine ones. The IBM Spyre Accelerator, with 32 dedicated AI acceleration cores per card, offloads high-volume real-time inferencing such as scoring a credit application, so it never competes with everyday processing.
“It’s about throughput,” says Christoforides. “Instead of burdening the normal Power processor with thousands of requests, you offload that inferencing to the Spyre cards and get the result a lot quicker.”
Alongside Spyre, IBM Bob, an AI-based developer tool, works out what old code is actually doing, the unglamorous discovery work that used to eat hours before any modernisation could begin. It analyses and documents programs, maps inputs and dependencies, and even generates test cases for QA and user acceptance. In one customer engagement, a six-hour developer task was completed in 20 minutes with IBM Bob.
Compliance by design
“Without the right tools, these are genuinely complex, multi-step tasks for any IT team, especially when something critical is happening,” says Christoforides. For regulated sectors like banking and healthcare, Power11's value isn't one feature. It’s lower licensing costs, reduced planned downtime, built-in AI, quantum-safe encryption, ransomware resilience and features that help support customers’ regulatory compliance obligations for mission-critical banking and payment systems. A consistent experience across on-premises and cloud deployments also removes a layer of complexity IT teams would otherwise have to manage themselves.
That is also where Anykey’s IBM expertise becomes practical, helping clients design maintenance and migration plans, structure the upgrade path and frame the AI conversation around their own environment rather than a generic one. “Customers are increasingly open to talking about AI infrastructure as they start that journey,” he ends. “The adoption is already under way. It’s now about how seriously organisations engage with what Power11 can actually do for them.”
Results may vary depending on configuration, workload and environment. This release may contain forward-looking statements regarding future products and capabilities, which are subject to change.
Share
