Hybrid Computing
-
Moving AI Workloads from Cloud to On-Premise: A Strategy for Reducing Power Consumption
Arm CEO Rene Haas advocates for distributing AI workloads from cloud-based infrastructure to local devices to reduce energy consumption and improve sustainability. He highlights a shift towards hybrid computing, with AI training in the cloud and inference occurring on devices like smartphones and AR glasses. Arm’s expanded partnership with Meta aims to optimize AI efficiency across the entire compute stack, exemplified by localized speech recognition in Meta’s Ray-Ban Wayfarer glasses. This localized processing enhances responsiveness and reduces reliance on cloud servers.
-
MediaTek Unveils Edge-to-Cloud AI Vision at COMPUTEX 2025
MediaTek introduced hybrid AI computing solutions at Computex 2025, including the 5G Generative AI Gateway for edge-cloud integration, NVIDIA collaborations (Genio IoT platform and NVIDIA GB10 Superchip for desktop-scale 200B model execution), and context-aware automotive systems (Dimensity Auto with C-X1 chip). Innovations also featured 6G-compatible connectivity in wearables, Ku-band satellite modems, and SMR100 mini-LED displays with 276% higher brightness than OLED. These advancements target unified, high-performance AI accessibility across decentralized infrastructures, enterprise, and consumer sectors.