CE LAB to Join Ministry of Science and ICT's 'AIDC Alliance'

CE LAB, a company specializing in artificial intelligence (AI) infrastructure software, is working to strengthen its engineering capabilities for GPU infrastructure.

On July 29, CE LAB announced that it would participate in the AIDC Alliance, a public-private consultative body led by the Ministry of Science and ICT.

The AIDC Alliance is a consultative group established jointly by the government, industry, academia, and research institutes to foster AI Data Centers (AIDC) as a national strategic industry and a new export sector. It serves as a forum for discussing issues and policy improvements related to AIDC construction, including power supply, land, and finance. The Ministry of Science and ICT is aiming to launch the Alliance in the second half of this year.

The government plans to build a total of 18.4GW of AIDC capacity by 2035, expand cumulative investments to over 1,000 trillion won, and foster a 'full-stack ecosystem' that combines domestic AI semiconductors, cloud platforms, power and cooling solutions, and operating software.

Recently, AI data centers have been evolving into high-performance computing (HPC) clusters, operating thousands to tens of thousands of GPUs as a single system. Starting with NVIDIA's next-generation GPU ‘Vera Rubin,' compliance with reference architecture (RA) and implementation of rack-level integrated systems has become a key requirement.

CE LAB possesses professional service capabilities spanning the entire process, including cluster design, RA-compliant infrastructure, performance verification, and supply of operating software. These are based on its ten years of accumulated GPU system engineering know-how and its qualification as an ‘Elite’ partner, the highest level in NVIDIA’s software division.

In the operational stage after building the data center, the company maximizes the efficiency of idle GPU resources using its flagship solution, AstraGo, which supports Kubernetes-based GPU scheduling, automatic GPU partitioning, and multi-node distributed learning.

Jungwhan Chae, CEO of CE LAB, stated, “The success or failure of AI data centers depends on the level of engineering that maximizes GPU resource performance. Through participation in this alliance, we will provide engineering technology covering both construction and operation, contributing to strengthening the competitiveness of the domestic AI infrastructure ecosystem.”

This content was produced with the assistance of AI translation services.