Spyre Accelerator by IBM Now Commercially Available

IBM has announced the upcoming general availability of the IBM Spyre Accelerator, a next-generation AI accelerator designed to deliver low-latency inferencing for generative and agentic AI workloads while ensuring the security and resilience of core enterprise operations.

Initially revealed earlier this year, Spyre will be available starting October 28 for IBM z17 and LinuxONE 5 systems, followed by Power11 servers in early December.

The modern IT landscape is rapidly shifting from traditional logic-based workflows to agentic AI inferencing, which demands ultra-low latency and real-time responsiveness. Recognizing this, IBM identified the need for mainframes and servers capable of running AI models alongside high-performance enterprise workloads—without sacrificing throughput or reliability.

To meet this demand, organizations require AI inferencing hardware that supports both generative and agentic AI while ensuring the security, resilience, and integrity of mission-critical data, transactions, and applications. The IBM Spyre Accelerator fulfills this need, allowing clients to keep sensitive data on-premises, minimizing risk, and improving both operational and energy efficiency.

Built on IBM’s research-to-product pipeline, the Spyre Accelerator merges cutting-edge innovation from the IBM Research AI Hardware Center with robust, enterprise-grade engineering from IBM Infrastructure.

Originally developed as a prototype chip, Spyre evolved through rapid iteration and real-world cluster testing at IBM’s Yorktown Heights campus and in collaboration with the University at Albany’s Center for Emerging Artificial Intelligence Systems.

The IBM Research prototype has matured into a fully enterprise-grade product designed for deployment across IBM Z, LinuxONE, and Power systems. Now known as the Spyre Accelerator, it is a commercial system-on-a-chip featuring 32 accelerator cores and 25.6 billion transistors, built using 5nm node technology.

Each Spyre unit is housed on a 75-watt PCIe card, enabling scalable configurations—up to 48 cards in IBM Z or LinuxONE systems and 16 cards in IBM Power systems—allowing organizations to expand AI capabilities efficiently and seamlessly.

“One of our key priorities has been advancing infrastructure to meet the demands of new and emerging AI workloads,”

said Barry Baker, COO, IBM Infrastructure & GM, IBM Systems.

“With the Spyre Accelerator, we’re extending the capabilities of our systems to support multi-model AI – including generative and agentic AI. This innovation positions clients to scale their AI-enabled mission-critical workloads with uncompromising security, resilience, and efficiency, while unlocking the value of their enterprise data.”

“We launched the IBM Research AI Hardware Center in 2019 with a mission to meet the rising computational demands of AI, even before the surge in LLMs and AI models we’ve recently seen,”

said Mukesh Khare, GM of IBM Semiconductors and VP of Hybrid Cloud, IBM.

“Now, amid increasing demand for advanced AI capabilities, we’re proud to see the first chip from the Center enter commercialization, designed to deliver improved performance and productivity to IBM’s mainframe and server clients.”