GMI Cloud, a leading NVIDIA Cloud Partner, has announced a landmark $500 million investment to establish a new AI Factory in Taiwan. This state-of-the-art facility is designed to be a cornerstone of the region's AI infrastructure, empowering businesses to train and deploy sophisticated AI models at a scale previously unattainable. It represents a significant step in building sovereign AI capacity while leveraging advanced U.S. technology.
GMI Cloud is building a $500 million, 16-megawatt AI Factory in Taiwan to bolster regional AI infrastructure.
The facility is powered by 7,000 next-generation NVIDIA Blackwell Ultra GPUs for massive computational power.
It is engineered for large-scale AI inference, model fine-tuning, and multi-modal workloads.
The AI Factory will accelerate real-world AI deployments across cybersecurity, manufacturing, and data infrastructure.
Key partners include Trend Micro, Wistron, VAST Data, and TECO, showcasing immediate industry applications.
This initiative strengthens a trans-Pacific AI ecosystem, uniting U.S. innovation with Asian manufacturing scale.
This initiative sets a new standard for sovereign AI capacity, creating a powerful link between American AI infrastructure innovation and Asian manufacturing and deployment expertise. The GMI Cloud Taiwan AI Factory embodies a model of cross-regional collaboration, focusing on accelerated computing, energy efficiency, and data sovereignty to power the next era of global AI.
The facility's core is powered by 7,000 NVIDIA Blackwell Ultra GPUs across 96 high-density racks, enabling it to process nearly 2 million tokens per second. Built on the full NVIDIA Blackwell architecture suite—including NVLink, Quantum InfiniBand, and Spectrum-X Ethernet networking—this 16-megawatt supercomputer is designed to transform AI infrastructure concepts into production reality. It will provide Taiwanese enterprises with the scalable, high-performance computing resources needed to accelerate advanced AI solutions.
The AI Factory's impact is already being realized through partnerships with industry leaders who are deploying its power for critical use cases. These collaborations demonstrate the practical application of AI factories across diverse sectors, moving beyond experimentation to industrialization.
Trend Micro is reinventing proactive security using digital twin technology accelerated by NVIDIA AI Enterprise and GMI Cloud's infrastructure. This allows organizations to simulate cyber threats and adapt policies in real-time without risking production systems.
Wistron is leveraging the AI Factory to build AI-enabled manufacturing systems. Using NVIDIA accelerated computing for computer vision and predictive maintenance, they are training and deploying models directly on production lines to reduce downtime and advance smart manufacturing.
VAST Data provides the high-performance data infrastructure foundation for the AI Factory. Its unified data architecture ensures exabyte-scale performance and resilience, delivering the high-throughput access required for continuous, large-scale AI workloads across thousands of GPUs.
TECO is utilizing the platform to pioneer Energy-as-a-Service solutions. By integrating its expertise in power systems with GMI Cloud and NVIDIA-accelerated computing, TECO is transforming its engineering knowledge into digital assets for optimizing energy in smart factories and modular data centers.
The launch of the GMI Cloud Taiwan AI Factory marks a pivotal moment in the industrialization of artificial intelligence in Asia. By uniting cutting-edge NVIDIA Blackwell technology with GMI Cloud's orchestration and regional expertise, this facility provides the sovereign infrastructure necessary for enterprises to transition from AI experimentation to full-scale production and innovation.
GMI Cloud delivers full-stack, U.S.-based GPU infrastructure and enterprise-grade inference services built to scale AI products. Whether training foundation models or deploying real-time agents, GMI gives teams full control of performance, costs, and launch velocity. With on-demand and reserved GPU clusters for all workloads and projects, GMI helps teams build AI without limits.