AI Beyond the Edge Forum: Edge AI Trends and Innodisk’s Evolution
Innodisk Group hosted the "AI Beyond the Edge" forum in Taipei in October 2025, convening global technology leaders and domain experts to explore the transformative potential of edge AI. The hybrid event drew approximately 1,200 attendees, both in-person and virtually, and attracted attention from numerous international media outlets.
The forum featured insightful discussions on how AI is moving beyond the cloud to the edge, enabling intelligent applications across manufacturing, transportation, healthcare, enterprise and more. Innodisk also unveiled its fully prepared strategic positioning and edge AI solutions together with partners from Axelera AI, Data Systems, Intel, KENMEC, NVIDIA and Qualcomm.
Chairman Randy Chien’s Opening Remarks
Innodisk’s Strategic Transformation into an “AI Architect
Innodisk Group Chairman Randy Chien opened the forum by analyzing the macro trends reshaping the industry. He observed that advances in generative AI and LLMs have fundamentally shifted enterprise conversations—the question is no longer whether to adopt AI, but how to implement it effectively at scale.
Chien also emphasized that the surge of generative AI is driving up model sizes and computing requirements, placing new demands on hardware. Innodisk aims to address client’s challenges by solving the most complex integration problems “before the last mile” of AI deployment.
Originally known as a specialist in industrial memory and storage, Innodisk has transformed itself into an “AI Architect.” To support this direction, the company has upgraded its product portfolio into nine modular AI Building Blocks and Intelligence that enable rapid solution development: Data Storage, I/O Extended, Peripheral Connecting, Remote Management, CPU/GPU/ASIC Computing, Sensor, AI Stack SDK, Partner Collaboration, and Applied Integration.
By breaking AI deployment into these fundamental blocks, Innodisk can mix and match components and solutions to meet diverse industry needs and accelerate the transition from proof of concept to large-scale deployment. With the integration complexity handled upfront, customers can focus on developing their AI applications instead of managing system integration.
Keynote Speakers and Highlights
Eason Hung, Senior Solution Architect, NVIDIA
The AI Revolution Paving the Way for Future Robotics Technology
Hung introduced NVIDIA's Physical AI vision—where AI systems not only process text and images but sense and act in the physical world. Evolving from text tokens to image tokens and ultimately action tokens, Physical AI will fundamentally reshape our physical environment.
NVIDIA－The AI Revolution Paving the Way for Future Robotics Technology
James Su, Director of Software, Aetina
Edge-First AI: The Strategic Blueprint for Enterprise and Industry Intelligence
Su highlighted the shift from cloud-centric AI to intelligent edge systems. He cautioned against chasing TOPS metrics alone, emphasizing the need to balance performance, responsiveness, and power consumption. Aetina delivers multi-architecture platforms and software-driven solutions that make Edge-First AI a deployable reality.
Aetina Corporation－Edge-First Al: The Strategic Blueprint for Enterpriseand Industry Intelligence
Evgeni Gousev, Senior Director, Qualcomm
Scaling AI for Everyone: Democratizing Intelligence at the Edge with Qualcomm Dragonwing
Gousev noted that over 75% of data is now generated locally, making Edge AI the "new oil." He introduced the Qualcomm Dragonwing product family and envisioned pervasive AI fully owned and controlled by enterprises at the edge.
Qualcomm－Scaling Al for Everyone: Democratizing Intelligence atthe Edge with Qualcomm Dragonwing
Ian Huang, Vice President of Manufacturing Business Group
Data Systems – Data-Intelligent Transformation for Innovation and Productivity
Huang addressed bridging the gap between productivity demands and labor capacity through AI. He presented Data Systems' multi-agent model (Commander, Expert Agent, Avatar Agent) that integrates AI into ERP, CRM, and business operations.
Data Systems－Data-Intelligent Transformation for Innovation and Productivity
Fabrizio Del Maffeo, CEO, Axelera AI
Axelera AI－From Edge to Everywhere: Scaling AI Beyond the Edge
Del Maffeo emphasized three essentials for edge AI success: high compute density, low cost, and low power. Axelera's Metis platform uses Digital In-Memory Computing, while the new Europa accelerator targets generative AI at the edge—all in service of democratizing AI.
Axelera AI－From Edge to Everywhere: Scaling AI Beyond the Edge
Wei Yeang Toh, General Manager of Ecosystem Development, Intel
Ecosystem Synergy: Unlocking New Global Opportunities in Edge AI
Toh shared Intel's 45-year edge computing commitment and 4,000+ partner ecosystem. He stressed that no company succeeds alone, highlighting Innodisk's APEX-E100 powered by Intel as an example of successful collaboration, and invited the industry to join forces.
Intel Corporation－Ecosystem Synergy: Unlocking New Global Opportunities in Edge Al
Intel Corporation－Ecosystem Synergy: Unlocking New Global Opportunities in Edge Al
Jackson Chen, Sales Director, KENMEC MECHANICAL ENGINEERING
Unlocking the Value of System Integration with AI, Digital Twin and Physical AI
Chen showcased how KENMEC combines AI, digital twins, and robotics to build a "system integration ecosystem with complete learning capability." Using Aetina MGX edge AI servers, they've created on-premises digital twin platforms and universal models for cross-robot operations.
KENMEC－Unlocking the Value of System Integration with AI, Digital Twin and Physical AI
Johnny Wu, Senior Director of Intelligent Peripheral Appliance Div., Innodisk Building Scalable AI Systems and Intelligent Infrastructure Architecture
Wu reiterated the Innodisk’s approach to transforming AI potential into scalable and deployable solutions.
Innodisk’s unique value proposition lies in enabling autonomy at the edge by providing complete peripheral support and end-to-end hardware–software integration.
He outlined four core elements of Innodisk’s intelligent edge architecture:
1. Sensing – With more than 80 percent of AI data coming from images, sensing is the foundation of intelligent decision-making.
2. Storage – Acting as an active memory extension to alleviate GPU VRAM bottlenecks in large language models.
3. I/O – The essential bridge that converts AI-driven decisions into precise physical actions.
4. AI Software – Enabling the transition toward Agentic AI, where systems can plan, make decisions, and continuously self-improve.
Wu’s presentation effectively tied back to Innodisk’s transformation into an AI Architect, and the vision to “ARCHITECT INTELLIGENCE”, highlighting its modular approach to AI deployment and echoing the earlier “building blocks” concept. He described Innodisk’s products as “building blocks that can be stacked and combined” to create complete, application-ready AI solutions.
Wu concluded by underscoring a message that resonated throughout the forum: “AI’s true value isn’t realized in the cloud, but at the edge — in those smart systems on-site that can operate, collaborate, and learn autonomously.”
Innodisk Corporation－Building Scalable AI Systems and Intelligent Infrastructure Architecture
Beyond the Stage
While keynote sessions explored the strategic and technological landscape of edge AI, the forum venue transformed into an interactive showcase. Exhibition booths featuring Innodisk and partner companies lined the conference hall, offering attendees hands-on access to cutting-edge products and live demonstrations of real-world use cases.
Visitors engaged directly with working prototypes, observed AI-powered solutions in action, and discussed implementation strategies with technical experts. This integration of thought leadership and tangible technology demonstrations provided attendees with both the strategic vision and practical insights needed to advance their own edge AI initiatives.
 |  |  |
Conclusion
The message across all keynotes was clear: AI is evolving from passive tools to autonomous systems capable of real-time judgment and collaboration at the edge. Whether through Physical AI, agentic systems, or humanoid robotics, the future demands intelligence deployed where action happens.
From cloud to edge to every device in between, Innodisk is building the intelligent infrastructure that powers this AI-driven future.
Notice:
This article includes images, video, and event-related content captured at public industry events for informational purposes. If you have any concerns regarding your appearance, statements, or referenced content, please contact us for review and appropriate action.
Further Reading
