PanaAI & Junee join forces for NVIDIA AI supercomputer

PanaAI has announced a partnership with Junee Ltd to construct the most powerful NVIDIA AI supercomputer in the Southern Hemisphere, which is set to be operational in Australia by the first quarter of 2025.

The new project, called "PanaAI AUS AISF", is designed to offer cloud services to both public and private sectors. It aims to facilitate the development of large foundational models for generative AI and support the advancement of research, innovation, and societal applications.

The PanaAI AUS AISF system plans to utilise up to 4,088 NVIDIA H200 Tensor Core GPUs interconnected through the NVIDIA Quantum-2 InfiniBand platform, addressing the rising computational demands of AI workloads.

The supercomputer's projected peak performance is approximately 16.4 exaflops in half-precision (FP8), making it potentially the largest AI supercomputing cluster in Australia and the fastest in the Southern Hemisphere.

PanaAI AUS AISF aims to tackle Australia's needs for significant computing resources to support complicated AI tasks in various sectors. The system's design includes NVIDIA's Hopper architecture-based H200 Tensor Core GPUs, which provide 1.8 times more memory and 1.4 times higher bandwidth compared to their predecessors. This enhanced capability allows for the management of larger model sizes and datasets, without the need for extensive data retrieval from slower storage, nor relying on complex techniques like tensor or pipeline parallelism.

The H200 GPUs deliver up to a 45% performance improvement over previous models in specific AI workloads, such as Llama and GPT.

This advancement supports large-scale artificial intelligence projects while emphasising top-tier security.

The PanaAI AUS AISF will supply accelerated computing cloud services, offering public cloud access, AI compute power rentals, and subscriptions aimed at training large language models. It is expected to rank among the top 10 most powerful AI supercomputers globally based on the TOP500 rankings as of June 2024. It seeks to advance research and innovation across sectors like healthcare, energy development, financial services, autonomous driving, life sciences, and public services.

A spokesperson for PanaAI stated, "To meet the growing demand for generative AI and computing resources, PanaAI is planning to establish multiple world-class, powerful supercomputing systems to provide unparalleled accelerated services to users worldwide. We highly value our collaboration with NVIDIA to deliver AI cloud services and rapidly train large foundational models built on ever-expanding datasets of images, acoustics, and sensor information. We look forward to closer collaboration to support users around the globe."

PanaAI specialises in AI computing services, particularly in GPU capabilities for areas such as high-performance computing and machine learning. As an NVIDIA NCP partner, PanaAI commits to ensuring that innovators and developers have the necessary computational resources to drive the future of AI.