IBM AI Infrastructure for IBM i
What Power10, Power11, and Spyre are built to run
Power10 and Power11 include Matrix Math Acceleration in every processor core. Spyre adds dedicated inference capacity to supported Power11 configurations, but it also brings Linux, Red Hat AI software, expansion hardware, and a separate support stack. Compare the complete system, not just the card.
Match IBM AI workloads to the infrastructure
Compare processor acceleration, supported Linux runtimes, data movement, and capacity before selecting a Power system or Spyre configuration.
AI workload placement
Compare cloud, an existing Power system, and a Linux inference partition before assuming AI must run inside IBM i.
Power S1112 local inference
IBM's one-socket Power11 announcement puts on-chip MMA and adjacent Linux partitions in the conversation. Check the workload fit.
Spyre Accelerator prerequisites
A PCIe-attached accelerator for supported Power11 configurations, with Red Hat and expansion-drawer requirements. It is not an IBM i-native card.
Infrastructure readiness
Inventory system headroom, partitions, data movement, storage, resilience, and licensing before a pilot.
IBM AI silicon specifications
Published figures for the Spyre Accelerator, on-chip Matrix Math Acceleration, Telum II, the AIU, NorthPole, and the 0.7 nm NanoStack research node.
The runtime the card requires
IBM names Red Hat AI Inference Server or Red Hat OpenShift AI as required for Spyre on Power11. Here is that stack, plus the open Granite models it serves.
IBM AI hardware news that changes a real decision
No press-release pile. Each report explains what changed, what it means for an IBM i shop, and what to check next.
IBM Made Spyre Support a One-Call Problem, but Only If You Buy the Stack Together
The best new Spyre feature is not on the card. IBM now says one Supportline case can start the repair across the accelerator, firmware, operating system, management software, and inference software when the stack is bought through IBM.
IBM Gave Power10 a Supported AI Starting Point Without a Spyre Card
IBM's Open-Source AI Foundation brings Power-optimized frameworks and services to Power10 and Power11. That gives an existing Power10 shop a real place to start before anyone turns the project into a new-hardware request.
IBM Opened the Spyre Card, and the Design Explains the Product
IBM Research has shown the card beneath the product sheet. Its local memory, accelerator cores, and tight power controls explain why Spyre is built for steady inference beside enterprise data rather than model training.
When another specialist site is the better answer
Hardware, software, and storage overlap in real projects. The coverage should not.
Software tools and IBM Bob
The watsonx family, IBM Bob, MCP connectivity, governance, and product comparisons belong on our software AI sibling.
FlashSystem AI storage
FlashSystem.ai and model-level storage detail belong on the dedicated FlashSystem site.
Power 11 system detail
For system specifications and generation-level comparisons, use the Power 11 specialist site.