The processor supports up to 64KB L1 and 512KB L2 caches, optimizing memory access performance.

The Arm Cortex-A320 is an ultra-efficient Armv9 CPU designed specifically for IoT, enhancing performance, efficiency, and security for edge devices.

While no public pricing or early design win details were provided, Arm expects the Cortex-A320 to be available in silicon by 2026.

Paul Williamson of Arm emphasized that this platform represents a significant advancement for edge computing, enabling intelligent decision-making closer to data collection points.

The new architecture includes advanced security features and AI enhancements, such as Scalable Vector Extensions (SVE2) and support for the BFloat16 data type.

The Cortex-A320 is designed to run lightweight AI workloads independently from the cloud, addressing the need for local data processing in applications like smart cities and industrial automation.

The platform integrates the Cortex-A320 CPU and Ethos-U85 NPU, capable of handling AI models with over one billion parameters on-device.

Arm's Kleidi AI platform enhances performance by up to 70% on the Cortex-A320 CPU when running specific datasets, optimizing ML and AI workloads.

The Cortex-A320 can be paired with the Ethos-U85 embedded neural processing unit (NPU) accelerator in scalable system-on-chip designs, accommodating diverse performance needs.

Key features of the Cortex-A320 include SVE2 technology for improved machine learning performance, offering up to ten times the performance of its predecessor, the Cortex-A35.

This CPU is versatile, suitable for applications ranging from smart devices to autonomous vehicles and utility robots, extending beyond traditional Cortex-M use cases.

AWS has integrated the new hardware into its IoT Greengrass platform, facilitating the development of modern edge AI applications such as anomaly detection in agriculture and smart manufacturing.