On May 27, 2024, Foxlink Group (2392) announced a strategic collaboration with Japan’s Ubitus K.K. and its energy leader brand Shinfox Energy (6806) to establish Ubilink.AI CO., Ltd. (Ubilink). This partnership aims to build a revolutionary AI computing center in Taiwan, significantly enhancing Taiwan’s global competitiveness in computational power.
In under six months, as of early November 2024, Ubilink has achieved rapid construction of Taiwan’s largest AI green-energy computing center, “Ubilink.” The center is equipped with 1,024 NVIDIA H100 GPUs and a Quantum-2 InfiniBand network platform, reaching a peak computational capacity of 45.82 PFLOPS (floating-point operations per second). Ubilink has formally applied to the Top500 List of global computing power rankings and is anticipated to rank within the top 20-22 worldwide, marking a groundbreaking achievement for Taiwan in the AI technology wave.
Wu-Hung Chen, General Manager of Ubilink, confidently stated that the Ubilink Computing Center (45.82 PFLOPS) far surpasses Taiwan’s national research institute’s TAIWANIA 2 (9 PFLOPS) and TAIWANIA 3 (2.3 PFLOPS), as well as NVIDIA’s Taipei-1 in the Kaohsiung Software Park (22.3 PFLOPS). As Taiwan’s largest AI supercomputing center, Ubilink will meet the robust computational demands for AI model training and inference across various industries, setting a new milestone in Taiwan’s AI computing capabilities.
Chen highlighted that the emergence of generative AI has sparked a tremendous demand for model training across sectors. Major Taiwanese IC design companies have already pre-booked computational resources for AI-driven IC design models. From IC design upstream to AI packaging and testing downstream, the demand for AI model testing within the semiconductor supply chain is an unstoppable wave.
In biopharmaceutical R&D, the increased precision of AI-driven computing has drastically shortened testing and development timelines, presenting revolutionary growth potential for the industry. Even in clinical diagnostics, hospitals will soon leverage computational automation to improve the accuracy of CT scans and endoscopic diagnostics, like liver and colonoscopy screenings. This advancement will greatly enhance healthcare quality, improving diagnostic precision and contributing significantly to patient care.
Additionally, the rapid development of Industry 4.0 and AI-driven modeling—paired with automated production efficiencies—promises to boost productivity in machinery and computing, creating higher margins and revenue. This leap will elevate Taiwan’s international competitiveness, setting new standards in industrial value creation and economic contribution.
