📝 Summary
Universiti Teknologi Malaysia (UTM) has established Malaysia’s first ultra-high-speed university GPU supercomputer, a state-of-the-art facility powered by NVIDIA Blackwell architecture GPUs and Intel Xeon processors. The facility, funded by the Malaysian Government, serves as a shared national resource for collaborative research and elevates the broader Malaysian academic community in the global AI ecosystem. The supercomputer provides unprecedented computational power for AI research, training, and development, preparing a future-ready workforce equipped with hands-on experience on frontier AI hardware.
KUALA LUMPUR, Oct 2 – Universiti Teknologi Malaysia (UTM) has officially established its state-of-the-art High-Performance Computing (HPC) GPU Infrastructure at the Faculty of Artificial Intelligence (FAI), launching the first ultra-high-speed Graphics Processing Unit (GPU) supercomputing facility among Malaysian universities. This landmark facility is made possible through the funding provided by the Malaysian Government, as part of the RM20 million investment allocated by the Ministry of Higher Education (MOHE) to UTM under Budget 2024 to establish Malaysia’s first AI faculty and strengthen the nation’s AI talent ecosystem.
UTM expresses its sincere gratitude to MOHE for its trust and vision in establishing the FAI. This strategic trust has enabled the creation of a world-class HPC infrastructure that directly benefits students and researchers across UTM as well as other Higher Education Institutions (IPTs) nationwide. By serving as a shared national resource, the facility accelerates collaborative research and elevates the broader Malaysian academic community into a competitive regional leader in the global artificial intelligence (AI) ecosystem.

At the center of this facility is Malaysia’s first high-speed university GPU supercomputer, powered by enterprise NVIDIA Blackwell architecture GPUs featuring 1,440 gigabytes (GB) of High Bandwidth Memory 3e (HBM3e), dual Intel Xeon processors with over 100 central processing unit (CPU) cores, and 2 terabytes (TB) of high-speed Double Data Rate 5 (DDR5) Random-Access Memory (RAM). Generating an unprecedented peak performance of up to 72 Peta Floating-Point Operations per second (PFLOPS) for model training and 144 PFLOPS for inference, the cluster utilizes fifth-generation NVLink/NVSwitch interconnects and ultra-low-latency InfiniBand networking.
Backed by 30 TB of high-speed Non-Volatile Memory Express (NVMe) storage and 100 TB of shared storage, this enterprise server gives researchers the computational power required to train foundation models from scratch and serve complex, trillion-parameter Large Language Models (LLMs) in real time.

Complementing the central supercomputer is the UTM AI Innovation Lab, equipped with a massive fleet of 120 dedicated high-performance desktop AI workstations. Each unit features an Intel Core Ultra 7 processor and a Blackwell-powered NVIDIA GeForce RTX 5070 GPU with 1TB NVMe storage, creating a robust local edge tier. This dual-tier setup ensures a seamless development pipeline, allowing 120 students and researchers to simultaneously build, debug, and prototype deep learning algorithms locally before deploying large-scale workloads directly to the primary supercomputer cluster without resource queuing.
This hybrid infrastructure provides a comprehensive, end-to-end framework for the entire AI research lifecycle, from initial model design and computer vision experiments to fine-tuning massive parameter models on domain-specific datasets. By combining localized workstation access with a centralized supercomputing core, UTM empowers researchers from UTM and partner IPTs to address complex global challenges across critical sectors including healthcare, autonomous systems, smart manufacturing, and natural language processing.

The deployment of this world-class GPU infrastructure underscores UTM’s commitment to driving cutting-edge technological advancements and supporting Malaysia’s national digital transformation agenda. By offering unprecedented computational power to both undergraduates and postgraduates, UTM is preparing a future-ready workforce equipped with direct, hands-on experience on frontier AI hardware, cementing its role as a leading research institution in Asia.
By: Dr. Mohd Syahid Mohd Anuar and Prof. Ts. Dr. Mohd Naz’ri Mahrin, Faculty of Artificial Intelligence, UTM