With the rapid advancement of artificial intelligence (AI) technologies, industries around the world are undergoing a new wave of structural transformation. AI is not only changing digital application models, but also driving rapid growth in infrastructure demands such as data centers, power systems, backup architectures, and cooling technologies. To explore emerging industrial trends in the age of AI, the College of Business, in collaboration with the Office of Academic Affairs and the Global Taiwanese Business Research Center, jointly organized the forum titled “Smart Transformation: AI-Driven Industrial Innovation Forum.” Industry experts were invited to share insights on the development and application of key infrastructure technologies in the AI era, helping enterprises seize opportunities for industrial upgrading driven by AI.
In the opening remarks, Professor Lin Chen-Yen from the Department of Business Administration at Chung Yuan Christian University, who also serves as Director of the Global Taiwanese Business Research Center, noted that society has often been described as “M-shaped,” referring to the shrinking middle class and expansion at both ends. However, under the wave of AI, industrial and social development is increasingly exhibiting a “K-shaped” pattern. This means that some enterprises are experiencing rapid growth by leveraging new technologies, while others risk marginalization due to failure to transform. Professor Lin further explained that AI is not a single industry but a comprehensive ecosystem, spanning from materials and equipment, semiconductor chips, and data center infrastructure to AI platforms and various application layers, forming a multi-layered industrial chain. He also cited NVIDIA founder Jensen Huang’s “Five-Layer AI Cake” theory, which categorizes the AI industry into power, chips, infrastructure, platforms, and applications. Among these, power supply, backup, and cooling systems for AI data centers are critical foundations supporting the entire AI ecosystem. He concluded with a humorous pun: “Without AI, there will be BI (悲哀, meaning ‘sadness’),” reminding both enterprises and individuals to actively connect with the AI ecosystem in order to stay on the upward trajectory in this K-shaped development era.
The first session of the forum focused on “AI Power Revolution: Innovations in Power Management Architecture for High-Performance Computing.” Mr. Chen Ying-Yuan, Vice President and General Manager of the Power and System Business Group at Delta Electronics, the second-largest listed company in Taiwan by market capitalization, shared insights on the challenges posed by the rapid expansion of AI data centers. He noted that the total power consumption of AI server racks may exceed 1 MW in the future, presenting unprecedented challenges to power infrastructure. In response, Delta Electronics has proposed three key technological directions: introducing high-voltage direct current (HVDC) architecture to improve power supply efficiency, developing 800V DC systems to support next-generation liquid-cooled servers, and integrating liquid cooling and two-phase cooling technologies to reduce energy consumption and improve Power Usage Effectiveness (PUE). With the rise of generative AI and high-performance computing, data centers are demanding significantly higher levels of efficiency and stability in power systems. Delta is actively developing high-power AI server power systems to meet the computing demands of next-generation AI data centers.
The second session, titled “AI Backup Revolution: Strategic Deployment of BBU for High-Stability Computing,” featured Mr. Huang Shih-Ming, Chairman of AcBel Polytech Inc., a company specializing in battery modules and energy solutions. He discussed power backup technologies for AI data centers operating under high-load computing environments. As the computational density of AI servers and data centers continues to rise, power stability has become a core element of AI infrastructure. Traditionally, data centers have relied on UPS (Uninterruptible Power Supply) systems for backup power. However, in high-density AI computing environments, UPS systems are becoming less suitable due to their large size, lower efficiency, and multiple power conversion layers. Battery Backup Units (BBUs) are emerging as a key solution for AI server backup power. Typically integrated directly into servers or racks, BBUs can provide immediate power support during fluctuations or outages, ensuring uninterrupted AI operations while reducing power conversion losses and improving overall energy efficiency. As demand for AI computing continues to grow, BBU systems not only enhance power stability but also reduce energy costs. In the future, traditional UPS systems are expected to be gradually replaced by BBUs, forming a new backup power architecture in AI data centers.
The third session, “AI Cooling Revolution: Innovations in Cooling Technologies for High-Density Computing,” featured Mr. Lee Wang-Rui, General Manager of the PBU division at Fushida, whose parent company is the thermal solutions leader Auras Technology. Mr. Lee pointed out that as AI chip computing power rapidly increases, data center cooling demands are experiencing exponential growth. In the past, servers primarily relied on fans and air-cooling systems. However, in high-density AI computing environments, the power consumption of individual GPU chips has surged, making traditional air cooling insufficient. As a result, cooling technologies have evolved from auxiliary components into critical systems that sustain AI data center operations. To address the high power and density requirements of AI servers, the industry is rapidly adopting liquid cooling and microchannel cooling technologies. Through high-precision microfluidic channel designs, cooling liquids can directly dissipate large amounts of heat generated by chips, significantly improving cooling efficiency while reducing energy consumption. As AI data centers continue to expand, cooling technologies have advanced from traditional mechanical design to highly integrated precision engineering, forming a crucial pillar of AI infrastructure alongside power and backup systems.
Chung Yuan Christian University emphasized that competition in the AI era is no longer limited to algorithms and applications, but has extended to infrastructure capabilities such as power, backup, and cooling. Through this forum, the university aims to promote deeper exchanges between academia and industry on the AI ecosystem, assist enterprises in seizing opportunities for industrial transformation, and cultivate interdisciplinary AI talent. This will also enable business students to gain a deeper understanding of the rapid development of the AI industry. The university further hopes that by continuing to host similar forums and industry-academia collaboration activities, it can strengthen the application of smart technologies and industrial partnerships, helping Taiwanese enterprises seize new opportunities in the global AI competition.