At COMPUTEX 2025, Arm outlined its expanding role in AI infrastructure, highlighting growth across cloud, edge, and device ecosystems. Chris Bergey, SVP and GM of Arm’s Client Line of Business, emphasized how Arm’s system-level approach and energy-efficient architecture are enabling scalable AI compute across a wide range of platforms—from datacenters to smartphones. Arm reported that nearly half of new server chips shipped to top hyperscalers this year are Arm-based, with adoption accelerating at AWS, Google, and Microsoft.
Arm also highlighted its growing footprint in AI PCs, smartphones, and Chromebooks, supported by its Compute Subsystems (CSS) for Client platform and the Armv9 architecture. MediaTek’s Kompanio Ultra SoC, built on Arm IP, was featured for bringing AI and multimedia capabilities to Chromebook Plus devices. NVIDIA’s Grace CPU and DGX Spark desktop platforms, also based on Armv9, were cited as examples of how Arm technology is enabling high-performance AI workloads for developers and researchers, with support from system manufacturers including Acer, ASUS, Dell, HP, Lenovo, and MSI.
The keynote also included updates on Arm’s next-generation Travis CPU and Drage GPU, designed to enhance AI acceleration and multimedia experiences in future devices. On the software front, Arm’s Kleidi AI library has now been integrated with major frameworks including PyTorch, llama.cpp, and ONNX Runtime, achieving more than 8 billion cumulative installs across Arm-powered devices. The company emphasized that AI inference, particularly at the edge, will continue to drive new hardware and software requirements as models scale and adoption grows.
“With the Arm compute platform at the heart of this transformation from cloud to edge, we’re not just enabling AI everywhere now – we’re shaping its future,” said Chris Bergey, SVP and GM, Client Line of Business, Arm.