Arm Holdings Limited (NASDAQ: ARM, hereinafter referred to as Arm) recently announced the expansion of the Arm® Flexible Access program to include the world’s first Armv9 edge AI computing platform optimized for IoT and edge AI workloads. This platform comprises the Arm Cortex®-A320 CPU and the Arm Ethos™-U85 NPU, both of which will be integrated into the Arm Flexible Access program in November 2025 and early 2026, respectively. This move will further inject rapid innovation momentum into the entire ecosystem, enabling both small startups and leading global OEMs to accelerate the development of the next generation of intelligent edge devices.
Paul Williamson, Senior Vice President and General Manager of Arm’s IoT Division, stated: “The next wave of AI computing innovation will unfold at the edge, thriving in devices, interfaces, and systems that bring intelligence closer to the source of data generation. From wearables and AI assistants to smart cameras, smart home devices, and robotics, the realization of these scenarios and true opportunities hinges on having high-performance, scalable computing capabilities.”
To assist partners in truly seizing this opportunity, Arm is continuously lowering the technical barriers to access, allowing every company in the ecosystem to more easily obtain Arm’s leading computing technologies, significantly speeding up product time-to-market.
Lowering barriers and accelerating innovation: The Arm Flexible Access program introduces the Armv9 edge AI computing platform.
With the introduction of the Armv9 edge AI computing platform into the Arm Flexible Access program, Arm will provide all companies dedicated to developing AI-capable edge devices with a smoother market entry path, a more competitive cost structure, and a more efficient innovation channel. This platform not only achieves the high performance and energy efficiency required for the large-scale deployment of intelligent edge devices but also incorporates security mechanisms to ensure device safety.
Launched earlier this year, the Armv9 edge AI computing platform features the ultra-energy-efficient Arm Cortex-A320 CPU and Arm Ethos-U85 NPU, capable of supporting edge AI models with over 1 billion parameters; it also integrates advanced Armv9 security technologies such as Pointer Authentication Code (PAC), Branch Target Identification (BTI), and Memory Tagging Extension (MTE) to protect critical applications and data at the edge.
Combining the machine learning (ML) performance of the Armv9 architecture, such as the second-generation Scalable Vector Extension (SVE2), with the energy efficiency advantages of the Ethos-U85, the Armv9 edge AI computing platform lays a solid foundation for the next generation of edge AI applications. Whether for smart cameras, smart home applications, industrial automation, or advanced human-machine interaction interfaces, it enables smarter and more natural interactions through vision, voice, and gestures, allowing machines to possess human-like perception and response capabilities while ensuring that inference and data processing are securely completed at the edge.
The expansion of the Arm Flexible Access program is an important milestone in driving edge AI innovation, fully demonstrating Arm’s firm commitment to ecosystem development. According to VDC Research, by 2028, AI will become the mainstream application technology for IoT projects, and the Arm computing platform is already at the core of this transformation.
With the Arm Flexible Access program, partners can access Arm’s extensive computing technologies, tools, and resources at low or no cost (for eligible startups) in the early stages. This model has become a significant catalyst for industry innovation, allowing companies to freely conduct design experiments and iterations, only needing to pay licensing fees for the technologies used in the final design. Through this model, partners such as Raspberry Pi, Weeteq, Hailo, and SiMa.ai have successfully developed groundbreaking edge AI technologies. Over the past five years, approximately 400 successful tape-outs have been achieved through the Arm Flexible Access program, with over 300 active members. More than half of these designs are aimed at embedded and IoT applications, injecting new intelligent power into edge intelligence.
Defining the Edge AI Computing Landscape
Arm is committed to promoting the widespread adoption of leading global technologies, empowering every participant in the ecosystem with innovative capabilities. With the introduction of the Armv9 architecture into the Arm Flexible Access program, Arm is providing the next generation of innovators with advanced architectures and supporting tools for edge AI, helping them meet the growing market demand for higher performance, enhanced AI capabilities, better energy efficiency, and stronger security.