Edge AI in Action - Industry Panel on Real-World Business Applications

Explore how industry leaders are transforming AI deployment from cloud-dependent to edge-optimized solutions in this 43-minute panel discussion from EDGE AI Taipei 2025. Discover how chipmakers are evolving into platform companies by combining scalable TOPS with comprehensive development ecosystems including compilers, model hubs, and ready-to-deploy containers that accelerate integration and deployment cycles. Learn from Renesas about managing hardware complexity through consistent APIs and multi-frontend support for TensorFlow Lite, PyTorch, ExecuTorch, and ONNX frameworks. Understand Qualcomm's developer-first methodology featuring optimizations, reference applications, and complete toolchains that bridge the gap from prototype to production deployment. Examine Advantech's approach to simplifying rollouts using prebuilt containers and over-the-air integration capabilities that eliminate toolchain fragmentation. Analyze Enerzai's successful deployment of sub-100 MB voice LLMs across 2 million set-top boxes, achieving cloud-quality performance while dramatically reducing latency and operational costs. Investigate Kudrat AI's implementation of on-device detection systems that extend battery life, protect privacy, and scale effectively across remote ecosystems. Address critical industry challenges including operator support requirements, embedded machine learning talent shortages, and the development of pragmatic standards that ensure efficient model execution rather than fallback scenarios. Gain insights into the strategic shift from cloud-versus-edge thinking to intelligent AI placement that optimizes cost, performance, and privacy considerations at scale.