Siemens has collaborated with Arm to help verification of the Arm® AGI CPU and validate its efficiency for next-generation agentic AI workloads, enabling scalable, production-ready infrastructure.
The Arm AGI CPU, constructed on the Arm Neoverse® Compute Subsystem (CSS) V3 platform, is designed to ship high-performance, energy-efficient compute for agentic AI and cloud knowledge heart deployments. Siemens’ Veloce™ Strato CS hardware-assisted verification platform was used to help verification of the Arm AGI CPU from subsystem by full-system degree, serving to to handle efficiency, latency and energy targets essential to hyperscaler deployment previous to tapeout.
Arm leveraged Siemens’ hardware-assisted verification (HAV), emulation and prototyping options to handle the complexity and efficiency calls for of its Neoverse V-series Compute Subsystem (CSS), with workflows engineered in collaboration with Arm to help validation of key efficiency metrics for hyperscale deployments.
“The size and complexity of contemporary AI compute platforms demand an enhanced degree of verification,” stated Karima Dridi, vp of productiveness engineering, Arm. “Siemens’ options enabled full-system verification of the Arm AGI CPU at scale utilizing a number of towers of Veloce Strato CS, serving to to make sure it meets the efficiency and effectivity necessities of hyperscale deployments.”
The Arm AGI CPU represents a breakthrough in performance-per-watt for agentic AI and high-performance computing workloads. The complexity of multi-die CSS designs incorporating Neoverse V-series cores, high-speed interconnects, PCIe Gen6, NVME and CXL demanded verification at a scale and constancy that conventional EDA toolsets alone can not ship.
Whereas Veloce Strato CS supplied full-chip verification at scale, Siemens’ Veloce proFPGA CS prototyping platform accelerates pre-silicon software program growth. By delivering FPGA-based prototypes working at near-real-time speeds, Veloce proFPGA CS permits software program groups to start validation, driver growth, and system bring-up months forward of silicon availability – serving to to make sure software program readiness aligned with aggressive silicon schedules and de-risking essential hyperscaler deployment milestones.
“The Arm AGI CPU exemplifies the engineering ambition driving agentic AI innovation,” stated Jean-Marie Brunet, senior vp and common supervisor of {Hardware}-Assisted Verification, Siemens Digital Industries Software program. “Working with Arm helps to make sure Siemens’ verification know-how can confirm full Neoverse CSS subsystems in a full SoC mannequin – at industry-leading efficiency and channel bandwidth. This achievement displays our shared dedication to pushing the boundaries of what is potential in AI infrastructure and our options’ capability to scale with prospects throughout the Arm ecosystem as they deal with more and more complicated architectural challenges.”
Constructing an ecosystem-based on superior verification
Siemens’ verification and implementation options combine seamlessly into Arm CPU design workflows, enabling fast adoption with minimal disruption. Critically, the identical Veloce Strato CS applied sciences used to assist develop the Arm AGI CPU can be found to Arm licensees, SoC designers and ecosystem companions – creating an ecosystem that accelerates innovation, reduces threat, and permits extra predictable time-to-market throughout the broader agentic AI neighborhood. For hyperscalers constructing customized silicon on the Arm Neoverse CSS platform, entry comparable verification applied sciences gives a path to gigawatt-scale deployment confidence.
To be taught extra about Siemens and Arm dedication to innovation, technical collaboration and supporting the evolving wants of Arm and its ecosystem companions as they drive the way forward for agentic AI infrastructure and knowledge heart computing, go to: https://blogs.sw.siemens.com/hardware-assisted-verification/2026/04/30/the-million-dollar-bug-why-hyperscale-verification-cant-afford-delays/