MegazoneCloud, an AI and cloud company, said on June 25 it will participate as an integrated infrastructure operations operator in the “AI Computing Testbed Infrastructure Upgrade” project led by the Ministry of Science and ICT and the National IT Industry Promotion Agency (NIPA).

In the project, MegazoneCloud will be responsible for overall integrated operations of a domestic NPU-based cloud infrastructure. It will develop software for NPU resource allocation and integrated management and monitoring, link NPU allocation and monitoring software based on a multi-cloud environment, and carry out integrated operation and control of AI computing infrastructure and build a user environment.

The project is part of the “Early Commercialisation of AI Semiconductors and AX Demonstration Support” programme to be 추진되는 over three years from 2025 to 2027, and this year marks the second year. The second-year goals are to build an AI computing infrastructure with at least 60 petaflops of compute capacity, advance AI semiconductor compilers and cloud software, and demonstrate AI application services in healthcare and LLM fields. One petaflop is a unit meaning 1,000 trillion operations per second.

Other consortium participants include FuriosaAI, Rebellions, NHN Cloud, Hyperaccel and Naver Cloud. The project results are expected to be used to upgrade and expand cloud AI infrastructure, optimise and diversify AI models, and advance AI services and expand business models across industries including manufacturing, finance, healthcare, education and the public sector.

MegazoneCloud CRO In-cheol Hwang (황인철) said, “The competitiveness of commercialising domestic AI semiconductors must be supported by cloud platform capabilities that can operate and scale stably in real service environments.” He added, “We will accumulate experience in both domestic infrastructure 구축 and overseas on-site demonstrations to contribute to the practical maturation of the domestic AI semiconductor ecosystem.”