The Arm AGI CPU is based on the Arm Neoverse Compute Subsystem (CSS) V3 platform and is designed to provide high-performance, energy-efficient computing for cloud data centres and advanced AI applications. Siemens’ Veloce Strato CS hardware-assisted verification system was used to validate the design across both subsystem and full-system levels prior to tapeout.
The verification effort focused on meeting key performance, latency and power targets required for deployment in hyperscale environments. Arm utilised a combination of Siemens’ hardware-assisted verification techniques, including emulation and prototyping, to address the increasing complexity of modern AI compute systems.
“The scale and complexity of modern AI compute platforms demand an enhanced level of verification,” said Karima Dridi, vice president of productivity engineering at Arm. “Siemens’ solutions enabled full-system verification of the Arm AGI CPU at scale using multiple towers of Veloce Strato CS, helping to ensure it meets the performance and efficiency requirements of hyperscale deployments.”
The AGI CPU is intended to deliver improvements in performance per watt for both agentic AI and high-performance computing workloads. However, its multi-die architecture, which incorporates Neoverse V-series cores, high-speed interconnects and interfaces such as PCIe Gen6, NVMe and CXL, presents significant verification challenges that go beyond the capabilities of traditional electronic design automation tools.
To address these demands, Siemens’ Veloce Strato CS platform enabled large-scale full-chip verification, while its Veloce proFPGA CS prototyping system was used to support early-stage software development. The FPGA-based prototypes allow software teams to begin testing and system integration ahead of silicon availability, helping to align software readiness with development timelines.
Jean-Marie Brunet, senior vice president and general manager of hardware-assisted verification at Siemens Digital Industries Software, said the collaboration demonstrated the role of advanced verification tools in supporting next-generation AI systems. “The Arm AGI CPU exemplifies the engineering ambition driving agentic AI innovation,” he said. “Working with Arm helps to ensure Siemens’ verification technology can verify complete Neoverse CSS subsystems in a full SoC model – at industry-leading performance and channel bandwidth.”
Siemens said its verification and implementation tools are designed to integrate with Arm’s processor development workflows, allowing chip designers to adopt them with minimal disruption. The same technologies used in the AGI CPU project are also available to Arm’s broader ecosystem, including licensees and system-on-chip developers.
As demand for customised silicon continues to grow among hyperscale data centre operators, access to advanced verification platforms is expected to play an increasingly important role in accelerating development and reducing risk. The collaboration reflects wider industry efforts to support more complex AI infrastructure, where scalability and reliability are critical to large-scale deployment.