宜鼎集團AI Beyond The Edge: 深入剖析Edge AI生態系整合與應用

在 AI 巨浪中，Innodisk 不只是供應商，而是您的最佳合作夥伴。Innodisk 將以「智慧建築師」定位，剖析硬體、軟體、系統整合等各層優勢。從 Camera、I/O、SSD、DRAM，到 Edge Platform、管理工具與服務，以「設計」為核心，打造「高擴展性架構」（Scalable Architectures），串接出一條順暢的 AI 落地之路。

AI 正從雲端迅速擴散至邊緣，帶動企業營運從中央決策轉向現場即時智慧。邊緣 AI 結合即時處理與智慧推論，滿足低延遲、自主決策與隱私保護等需求，成為企業數位轉型關鍵動能。 然而，AI 真正落地仍面臨挑戰，Aetina 透過軟硬整合平台與多元合作生態圈，協助企業在效率與彈性間取得平衡。以「Edge-First」思維，從現場擴展至全局，打造可規模化的 AI 應用，實現營運全面升級。

AI innovation often starts where the constraints are toughest – at the edge, where power, cost, and latency leave no margin for inefficiency. In this keynote, Fabrizio Del Maffeo, CEO of Axelera AI, shares how the company’s Metis AI platform and Voyager SDK deliver high-performance, low-power inference that makes advanced computer vision possible in the most demanding edge environments. He will explore how this edge-first foundation naturally scales “beyond the edge” into larger, more complex deployments – from connected factories and smart cities to enterprise and data center workloads. Attendees will gain insight into Axelera AI’s technology roadmap, including the upcoming Titania platform, and learn how an edge-to-cloud approach can unlock new opportunities for efficiency, scalability, and accessibility in AI.

在 AI 蓬勃發展的時代，企業更應積極克服缺工、市場等快速變化，製造業也將迎來智慧轉型的關鍵轉捩點。本次將分享如何以 AI 強化企業韌性，讓 AI 不只內建、更能外擴，全面釋放創新生產力！

As edge AI continues to evolve, collaboration across the ecosystem is becoming essential to drive innovation and scalability. This presentation explores how synergy among system builders, software developers, and industry partners can unlock new global opportunities in edge AI. We will highlight how Intel is working closely with the ecosystem to deliver impactful solutions and drive more sustainable outcomes worldwide.

以 「AI、數位孿生、Physical AI」 為主軸，導入 AI Robot、數位孿生 AI 基礎建設、Physical AI 、AI Nose SLM 嗅覺語言模型等應用方案，並以 AI 驅動的智慧平台，提供智慧物流及半導體自動化物料搬運系統 R2S S2R 的 Close Loop 架構，一開始就做對決策，找出最佳解。

我們將探討邊緣智慧對機器人技術未來的影響，特別是在 AI Shift to the Edge 的背景下，如何成為自主型系統的關鍵契機。我們將介紹視覺語言模型（VLM）和大規模語言模型（LLM）在 Jetson 平台上的應用，展示即時畫面理解及自然語言控制的實現，並分析 NVIDIA 的 Isaac 生態系統和「Three Computers」架構如何優化自主系統設計。最後，我們將探討 NVIDIA Jetson AGX Thor 平台作為高效能邊緣 AI 平台的技術優勢，揭示未來機器人技術的發展趨勢。

As AI continues to evolve from centralized cloud deployments to distributed edge intelligence, the ability to scale inferencing across diverse environments—securely, efficiently, and affordably—is becoming mission-critical. This session explores how Qualcomm is driving this transformation by converging advanced AI capabilities with its Dragonwing Product Portfolio. From on-device AI across Dragonwing processors up to cloud-grade performance at the edge using the AI On-Prem Appliance, the Dragonwing portfolio of products and solutions combine heterogeneous compute, optimized AI software, and domain-specific acceleration to achieve exceptional performance per watt and per dollar—unlocking new possibilities for real-time intelligence in everything from smart cameras to industrial automation. Qualcomm’s AI On-Prem Appliance also serves as a powerful platform of choice for deploying Agentic AI at the edge—enabling autonomous, context-aware intelligence with low latency and high efficiency. Join us to discover how Qualcomm is working with the ecosystem to push the boundaries of edge-native AI.

在 AI 巨浪中，Innodisk 不只是供應商，而是您的最佳合作夥伴。Innodisk 將以「智慧建築師」定位，剖析硬體、軟體、系統整合等各層優勢。從 Camera、I/O、SSD、DRAM，到 Edge Platform、管理工具與服務，以「設計」為核心，打造「高擴展性架構」（Scalable Architectures），串接出一條順暢的 AI 落地之路。

AI 正從雲端迅速擴散至邊緣，帶動企業營運從中央決策轉向現場即時智慧。邊緣 AI 結合即時處理與智慧推論，滿足低延遲、自主決策與隱私保護等需求，成為企業數位轉型關鍵動能。 然而，AI 真正落地仍面臨挑戰，Aetina 透過軟硬整合平台與多元合作生態圈，協助企業在效率與彈性間取得平衡。以「Edge-First」思維，從現場擴展至全局，打造可規模化的 AI 應用，實現營運全面升級。

AI innovation often starts where the constraints are toughest – at the edge, where power, cost, and latency leave no margin for inefficiency. In this keynote, Fabrizio Del Maffeo, CEO of Axelera AI, shares how the company’s Metis AI platform and Voyager SDK deliver high-performance, low-power inference that makes advanced computer vision possible in the most demanding edge environments. He will explore how this edge-first foundation naturally scales “beyond the edge” into larger, more complex deployments – from connected factories and smart cities to enterprise and data center workloads. Attendees will gain insight into Axelera AI’s technology roadmap, including the upcoming Titania platform, and learn how an edge-to-cloud approach can unlock new opportunities for efficiency, scalability, and accessibility in AI.

在 AI 蓬勃發展的時代，企業更應積極克服缺工、市場等快速變化，製造業也將迎來智慧轉型的關鍵轉捩點。本次將分享如何以 AI 強化企業韌性，讓 AI 不只內建、更能外擴，全面釋放創新生產力！

As edge AI continues to evolve, collaboration across the ecosystem is becoming essential to drive innovation and scalability. This presentation explores how synergy among system builders, software developers, and industry partners can unlock new global opportunities in edge AI. We will highlight how Intel is working closely with the ecosystem to deliver impactful solutions and drive more sustainable outcomes worldwide.

以 「AI、數位孿生、Physical AI」 為主軸，導入 AI Robot、數位孿生 AI 基礎建設、Physical AI 、AI Nose SLM 嗅覺語言模型等應用方案，並以 AI 驅動的智慧平台，提供智慧物流及半導體自動化物料搬運系統 R2S S2R 的 Close Loop 架構，一開始就做對決策，找出最佳解。

我們將探討邊緣智慧對機器人技術未來的影響，特別是在 AI Shift to the Edge 的背景下，如何成為自主型系統的關鍵契機。我們將介紹視覺語言模型（VLM）和大規模語言模型（LLM）在 Jetson 平台上的應用，展示即時畫面理解及自然語言控制的實現，並分析 NVIDIA 的 Isaac 生態系統和「Three Computers」架構如何優化自主系統設計。最後，我們將探討 NVIDIA Jetson AGX Thor 平台作為高效能邊緣 AI 平台的技術優勢，揭示未來機器人技術的發展趨勢。

As AI continues to evolve from centralized cloud deployments to distributed edge intelligence, the ability to scale inferencing across diverse environments—securely, efficiently, and affordably—is becoming mission-critical. This session explores how Qualcomm is driving this transformation by converging advanced AI capabilities with its Dragonwing Product Portfolio. From on-device AI across Dragonwing processors up to cloud-grade performance at the edge using the AI On-Prem Appliance, the Dragonwing portfolio of products and solutions combine heterogeneous compute, optimized AI software, and domain-specific acceleration to achieve exceptional performance per watt and per dollar—unlocking new possibilities for real-time intelligence in everything from smart cameras to industrial automation. Qualcomm’s AI On-Prem Appliance also serves as a powerful platform of choice for deploying Agentic AI at the edge—enabling autonomous, context-aware intelligence with low latency and high efficiency. Join us to discover how Qualcomm is working with the ecosystem to push the boundaries of edge-native AI.