-

Korea Institute of Science and Technology Information Selects Hewlett Packard Enterprise to Build South Korea’s Largest Supercomputer

Korea’s sixth-generation supercomputer, KISTI-6, will be 100% direct liquid-cooled to enable AI and HPC simulation for researchers across academia and industry

SINGAPORE--(BUSINESS WIRE)--Hewlett Packard Enterprise (NYSE: HPE) today announced that it has been selected by the Korea Institute of Science and Technology Information (KISTI), a government-funded research institute, to build a more advanced and versatile computing architecture for its sixth national supercomputer, “KISTI-6”, which is expected to become the most powerful supercomputer in South Korea. The new supercomputer will feature a 100% fanless direct liquid cooling system architecture to deliver greater efficiency, sustainability, and density.

KISTI has made its supercomputing services accessible to a variety of users for many years - including research institutions, universities, and commercial companies of all sizes. The new supercomputer will serve a wide range of demanding data-intensive research projects and form the backbone of artificial intelligence (AI) and simulation research and development in Korea. It is expected to reach a theoretical peak performance of 600 petaflops, more than half an exaflop.

“As the leader in supercomputing, having built many of the world’s fastest and energy efficient systems for high performance computing (HPC) and AI, HPE is uniquely positioned to help organizations minimize power consumption while maximizing business outcomes,” said Joseph Yang, general manager, HPC and AI, APAC and India, Hewlett Packard Enterprise. “We are proud to support this strategic initiative to strengthen Korea’s capabilities in fundamental scientific research and advanced artificial intelligence, aligned with their broader mission to support national R&D advancement.”

“Our goal is to work with HPE, AMD and NVIDIA to cement Korea’s position as a leading country for supercomputing and sovereign AI innovation,” said Sik Lee, President, Korea Institute of Science and Technology Information. “We want to nurture experts in data utilization, high performance computing and AI technologies to help bolster our scientific and technological competitiveness. Making AI accessible and inclusive is critical to encouraging innovation across industries and enhancing societal impact.”

KISTI-6 will be built on the HPE Cray Supercomputing EX4000 system with two partitions, one featuring NVIDIA GH200 Grace Hopper Superchips and another featuring 5th Gen AMD EPYC™ processors. The latest accelerators will process the large amounts of data required for AI training, inference, and simulation. HPE Slingshot interconnect 400, an exascale-capable interconnect, will deliver fast data transfers between the compute and storage units at 400 gigabits-per-second speeds and allows workloads and applications to scale across the entire system.

“Researchers, scientists and professionals across the globe are increasingly relying on advanced and complex AI and simulation capabilities to meet the growing demands of supercomputing,” said John Josephakis, Vice President of Supercomputing and HPC at NVIDIA. “By collaborating with the Korea Institute of Science and Technology Information and HPE, we’re expanding access to high-performance computing, empowering industries and teams to innovate faster and more efficiently.”

“Leveraging our experience powering the world’s most advanced supercomputers, AMD is collaborating with HPE to deliver KISTI-6 using 5th Gen AMD EPYC™ processors – the industry’s leading CPU for AI,” said Peter Chambers, Managing Director of Sales, APAC and Country Manager, Australia, AMD. “With breakthrough performance and outstanding energy efficiency, we’re enabling researchers to solve complex problems fast and efficiently.”

With the increasing demand from data-intensive workloads, KISTI also selected HPE Cray Supercomputing Storage Systems E2000, integrating fast Non-Volatile Memory Express (NVMe) and high-capacity Near Line (NL)-SAS disks, alongside the HPE Data Management Framework (DMF), for efficient and scalable data management. The fully integrated solution, tailored to KISTI’s complex research requirements, will be designed and delivered by HPE and located in a government datacenter, specially adapted for liquid-cooled systems.

About Korea Institute of Science and Technology Information (KISTI)

The Korea Institute of Science and Technology Information (KISTI) is a government-funded national research institute dedicated to advancing science and technology through cutting-edge expertise in high-performance computing (HPC), big data, and research information services. With over 60 years of experience, KISTI is leading innovation in national R&D by harnessing the power of AI and HPC in the era of the Intelligent Revolution. For more information, visit: www.kisti.re.kr

About Hewlett Packard Enterprise

Hewlett Packard Enterprise (NYSE: HPE) is a global technology leader focused on developing intelligent solutions that allow customers to capture, analyze, and act upon data seamlessly. The company innovates across networking, hybrid cloud, and AI to help customers develop new business models, engage in new ways, and increase operational performance. For more information, visit: www.hpe.com.

AMD, the AMD logo, EPYC, and combinations thereof are trademarks of Advanced Micro Devices, Inc.

Contacts

Media Contact:
Harleen Sachdeva
harleen.sachdeva@hpe.com

Hewlett Packard Enterprise

NYSE:HPE

Release Versions

Contacts

Media Contact:
Harleen Sachdeva
harleen.sachdeva@hpe.com

More News From Hewlett Packard Enterprise

HPE Transforms Distributed AI Factories Into Intelligent AI grid Powered by NVIDIA

SAN JOSE, Calif.--(BUSINESS WIRE)--NVIDIA GTC 2026 – HPE (NYSE: HPE) today announced the HPE AI Grid, an end-to-end solution built on the NVIDIA reference architecture to securely connect AI factories and distributed inference clusters across regional and far‑edge sites. The HPE AI Grid enables service providers to deploy and operate thousands of distributed inference sites, turning AI installations into a single intelligent system. AI‑native applications require predictable, low‑latency, distr...

HPE Threat Labs Report Reveals Cyber Adversaries Are Morphing Their Business Model to Scale and Accelerate Attacks

HOUSTON--(BUSINESS WIRE)--HPE (NYSE: HPE) today unveiled the results of its inaugural cyberthreat research report, In the Wild, showing a striking shift in how modern cyber adversaries operate at scale across global industries and critical public sectors. Based on HPE’s analysis of live threat activity observed globally throughout 2025, the report shows that cybercrime has gone industrial, with attackers using automation and long-standing vulnerabilities to scale campaigns and repeatedly compro...

HPE Accelerates Secure, Scalable Production-ready AI Through New Innovations With NVIDIA

SAN JOSE, Calif.--(BUSINESS WIRE)--HPE (NYSE: HPE) today announced a significant expansion of the NVIDIA AI Computing by HPE portfolio, redefining how enterprises deploy, operationalize, and scale AI. Through its deep partnership and co-engineering with NVIDIA, HPE delivers an advanced portfolio of integrated and validated systems that speed time to value for AI while addressing scale, security, and governance requirements. “The AI race is fundamentally about speed, scale, and trust,” said Anto...
Back to Newsroom