Arm has announced the expansion of its Arm Flexible Access licensing model to include the world's first Armv9-based edge AI computing platform, optimized for Internet of Things (IoT) and edge AI workloads. Core components of this platform include the Arm Cortex-A320 CPU and the Arm Ethos-U85 NPU, which are expected to be officially included in the Flexible Access licensing portfolio in November 2025 and early 2026, respectively.
This design will further lower the barrier to entry for advanced AI technology, allowing the entire ecosystem, from small startups to large OEMs, to access Arm’s latest edge AI computing technology at a lower cost and more conveniently, accelerating the development of a new generation of intelligent edge devices.
Targeting the edge AI wave and lowering the threshold for technology acquisition
Paul Williamson, senior vice president and general manager of Arm's IoT business unit, said: "The next wave of AI computing innovation will begin at the edge...The real business opportunity lies in the high-performance, scalable computing power required to enable these use cases."
To help partners seize this opportunity, Arm has introduced its latest Armv9 edge AI computing platform into the Flexible Access licensing model, providing developers with a smoother market entry path and a more competitive cost structure.
Armv9 Edge AI Platform: Balancing Performance, Efficiency, and Security
The Armv9 architecture edge AI computing platform released earlier this year combines the ultra-energy-efficient Cortex-A320 CPU with the Ethos-U85 NPU, which can support edge AI models with over 10 billion parameters.
In addition to performance and efficiency, this platform also incorporates Armv9 architecture security technologies, including Pointer Authentication Code (PAC), Branch Target Identification (BTI), and Memory Tag Extension (MTE), providing hardware-level security protection for critical applications and data at the edge.
Combining the machine learning (ML) capabilities of the Armv9 architecture (such as the second-generation Scalable Vector Extensions (SVE2) instructions) with the NPU acceleration capabilities of the Ethos-U85, this platform lays the foundation for smart cameras, smart homes, industrial automation, and advanced human-machine interfaces (HMIs) that require natural interaction through vision, voice, and gestures, ensuring that AI inference and data processing can be performed securely on the device.
Flexible Access licensing model: 400 designs finalized, with over 300 members participating
The Arm Flexible Access licensing model allows customers to access a wide range of Arm IP assets, tools, and resources on an upfront, free, or low-cost basis (for qualified startups). Customers only pay for the technologies they implement when their final product design is finalized (taped out), significantly reducing upfront investment and experimental risk.
Arm emphasized that this model has successfully fostered innovation, helping partners including Raspberry Pi, Weeteq, Hailo, and SiMa.ai create groundbreaking edge AI technologies. Over the past five years, Flexible Access has enabled 400 design wins and boasts over 300 active members, with over half of these designs targeting embedded and IoT applications.
According to a VDC Research report, AI will become the dominant technology in IoT projects by 2028. Arm's introduction of the Armv9 architecture into the Flexible Access licensing model is a clear attempt to further solidify its core position in edge AI computing, providing the ecosystem with key tools for higher performance, stronger AI capabilities, greater efficiency, and improved security.
