Representative End User Clients
Representative Automation Clients
Representative Software Clients
The AI revolution is no longer confined to the cloud. As the world becomes increasingly connected and intelligent, from smart cities to industrial automation, the need to process AI workloads at the edge is essential. To meet this expanding need at the edge, Arm introduced the Armv9edge AI platform featuring the new Arm Cortex-A320 CPU and the Arm Ethos-U85 NPU edge AI accelerator, enabling the ability to run tuned large language models (LLMs) and small language models (SLMs) for agent-based AI applications AI models with over one billion parameters to run on-device.
OEMs are under pressure to meet growing computing demands across IoT applications, such as autonomous vehicles navigating factory floors, smart cameras that can adapt their functionality through software updates, and human-machine interfaces that offer more natural, AI-driven interactions. To innovate and scale at pace, flexibility is required to execute AI workloads where it makes sense, as well as more robust security and increased software flexibility. Arm’s new platform brings together a brand-new Armv9 CPU, Cortex-A320, along with the Ethos-U85 NPU with operator support for transformer networks, creating the first Armv9 edge AI platform optimized for IoT. The platform is said to deliver an 8x improvement in machine learning (ML) performance compared with the Cortex-M85-based platform launched by the company last year.
New Cortex-A320 for Next Generation Intelligent IoT Devices
The new Cortex-A320 brings advanced AI capabilities and developer benefits to IoT, extending the features of the Armv9 architecture to power-efficient devices. Cortex-A320 takes advantage of Armv9 architectural features, such as SVE2 for ML performance, and delivers a 10x ML performance uplift and 30 percent scalar performance uplift compared with its predecessor, Cortex-A35. The platform’s Armv9.2 architecture also brings advanced security features like Pointer Authentication (PAC), Branch Target Identification (BTI), and Memory Tagging Extension (MTE) to even the smallest Cortex-A devices.
Extending Arm Kleidi to IoT
One of the most significant barriers to edge AI adoption has been the complexity of software development and deployment and time-to-market. To address this issue, the company is extending its Arm Kleidi software ecosystem, a set of compute libraries for developers of AI frameworks designed to optimize AI and ML workloads on Arm-based CPUs with no additional developer work, to IoT. KleidiAI is already integrated into popular IoT AI frameworks, such as Llama.cpp and ExecuTorch or LiteRT via XNNPACK, accelerating the performance of key models, including Meta Llama 3 and Phi-3. As an example, Kleidi AI brings up to 70 percent more performance to the new Cortex-A320 when running Microsoft’s Tiny Stories dataset on Llama.cpp.
The new platform also maintains software compatibility with higher-performance Cortex-A processors. This scalability ensures that developers can build solutions that grow and adapt as requirements change. With access to the Armv9 ecosystem and compatibility with both Linux and real-time operating systems such as Zephyr, developers can leverage existing tools and knowledge and take advantage of software reuse, reducing time to market and lowering total cost of ownership.
Early partners for the new platform include Advantech, AWS, Eurotech, Siemens, and Renesas.
Learn more about Industrial AI's Role in Digital Transformation of the Manufacturing Industries.