South Korea is stepping up efforts to build the next generation of digital infrastructure for the artificial intelligence era, with researchers launching a new initiative to strengthen high-speed connectivity across the nation's growing AI ecosystem.

The effort focuses on developing technologies that improve inter-data-center connectivity while advancing technical standards and industry collaboration. The long-term objective is to create a national AI backbone capable of linking distributed computing resources into a unified platform.

As generative AI and large-scale foundation models continue to expand, the ability to share computing power and transfer massive volumes of data in real time has become increasingly important. As a result, technologies capable of delivering ultra-high bandwidth, ultra-low latency, and lossless transmission are emerging as critical components of AI infrastructure.

The Electronics and Telecommunications Research Institute (ETRI) announced the launch of the Hyper-service Interconnect (HINT) Forum to accelerate research, ecosystem development, and standards activities in this rapidly evolving field.

The forum was established to advance next-generation interconnect technologies across a broad range of domains, including AI data center networks, national core and backbone communications infrastructure, metro and access networks, and industrial convergence environments supporting smart manufacturing, autonomous mobility, and other emerging applications.

Connectivity Emerges as a New Driver of AI Competitiveness

The performance of AI systems is no longer determined solely by the power of GPUs or advanced semiconductors. Even the largest computing clusters can experience significant performance bottlenecks if data movement between processors and facilities cannot keep pace with growing workloads.

Global technology leaders are increasingly investing in optical interconnects, dedicated data-center links, AI-optimized switching systems, and advanced transport architectures alongside their AI infrastructure expansion plans. Companies such as NVIDIA, Google, Microsoft, and Meta increasingly view connectivity as a strategic asset in large-scale AI deployment.

Industry analysts now regard digital connectivity, alongside semiconductors and power systems, as one of the three pillars shaping national AI leadership.

Connecting Korea’s Distributed AI Capacity

ETRI researchers are developing a range of technologies, including ultra-low-latency, lossless, high-bandwidth optical packet transport systems between AI facilities, AI workload-driven open transport network fabric operating systems, and ecosystem technologies designed to support future interoperability and standards development.

The institute expects these efforts to strengthen South Korea’s AI infrastructure and improve its ability to support next-generation AI services and large-scale computing workloads.

ETRI is also participating in the Ministry of Science and ICT’s Data Center Network Infrastructure Leadership Program. The initiative includes research into technologies that enable horizontal expansion of AI resources across multiple facilities, optical-network-based scale-up and scale-out architectures, and validation environments for testing next-generation infrastructure equipment.

The project represents a shift beyond isolated data-center operations toward a model in which geographically distributed computing resources can operate as a single integrated platform.

Experts believe that as large-scale AI compute clusters become increasingly interconnected, countries will be able to build national-scale AI supercomputing environments capable of dynamically allocating resources across regions, industries, and applications.

Global Competition Expands Beyond Technology to Standards

Competition in AI connectivity now extends beyond technology development to the race for global standards.

The United States, China, and major European economies are actively pursuing leadership in AI networking architectures, optical transport technologies, and AI workload optimization frameworks. Establishing internationally recognized standards can provide long-term advantages in equipment markets, software ecosystems, and future infrastructure deployment.

Through collaboration among industry, academia, and research institutions, the HINT Forum aims to strengthen South Korea’s influence in global standards development while supporting domestic companies seeking opportunities in emerging AI infrastructure markets.

As AI competition evolves from a race centered on processors to a broader contest involving connectivity, power systems, and data-center operations, leadership in networking technologies is becoming a strategic priority for countries pursuing sovereign AI capabilities.

Expanding the Industry-Academia-Research Ecosystem

ETRI hosted the inaugural HINT Forum General Assembly and Workshop at Gonjiam Resort in Gyeonggi Province on June 11–12.

Approximately 90 experts from industry, academia, and research institutions attended the event. In addition to the forum’s official launch, participants held a kickoff meeting for the AI Data Center Network Consortium.

Attendees shared insights on technology trends, research and development programs, and standardization strategies while discussing ways to expand cooperation across the AI networking ecosystem.

The event also included a plaque presentation ceremony recognizing organizations designated by the Ministry of Science and ICT as specialized research centers for data-center networking.

Kang Tae-gyu, Principal Researcher at ETRI and Chair of the HINT Forum Organizing Committee, said that efficient connectivity would become as important as computing resources in determining national competitiveness in the AI era.

“Beyond computing power itself, the ability to connect resources efficiently will become a decisive factor in national competitiveness,” Kang said. “We hope the HINT Forum will establish itself as a leading platform for AI networking technologies and standards development in Korea.”

The HINT Forum is supported by the Telecommunications Technology Association (TTA) under the Ministry of Science and ICT’s ICT Standardization Support Program. It plans to promote the commercialization of domestic AI networking technologies while expanding participation in international standards activities.

The event was jointly organized by the HINT Forum and the AI Data Center Network Consortium, and co-hosted by the Telecommunications Technology Association (TTA) and the Open Systems Interconnection Association (OSIA).

Why It Matters

The AI race is increasingly shifting from a contest over raw computing power to one centered on how effectively nations can orchestrate and utilize distributed compute capacity. The launch of the HINT Forum signals South Korea’s ambition not only to develop advanced AI semiconductors, but also to build the digital backbone required to support the next generation of national-scale AI systems.

Yeon Choul-woong More by this reporter