South Korea's NAVER Signs AI Computing Infrastructure Agreement with Brookfield Worth Up to US$9 Billion

en.Wedoany.com Reported - On October 27, South Korea's NAVER signed an AI computing infrastructure agreement worth up to US$9 billion with Canada's Brookfield Asset Management to advance an AI computing project jointly developed with NVIDIA. Brookfield will procure infrastructure assets including GPUs and data centers through a special purpose company, with NAVER Cloud responsible for operations. The project is centered on the GAK Sejong hyperscale data center in Sejong City, South Korea, and plans to operate 55 megawatts of computing capacity in the first half of 2027, expanding to 200 megawatts by 2028.

Under the agreement, Brookfield plans to establish a special purpose company tentatively named Asset Co. to undertake asset investment and procurement for the computing infrastructure. The company will be entirely funded by Brookfield and will provide computing capacity to NAVER Cloud after acquiring the relevant facilities. NAVER stated in its October 27 announcement that, under the International Financial Reporting Standards applicable in South Korea, the special purpose company is expected to be included in NAVER's consolidated financial statements.

The two parties have settled on a cooperative structure that separates infrastructure asset ownership from cloud computing operations. NAVER and its cloud business subsidiary will be responsible for operating the AI computing facilities, while Brookfield will handle investment arrangements for assets such as data centers and GPUs. NAVER has also agreed to provide a guarantee of up to eight years for NAVER Cloud's related usage fee payment obligations, with the usage fee payment arrangement linked to business revenue.

The formal agreement also satisfies the precondition for NVIDIA's strategic equity investment in NAVER. NVIDIA plans to invest US$1 billion in NAVER through a third-party allotment, and NAVER will proceed with the corresponding capital increase procedures. NVIDIA will also participate in technology development around high-performance computing platforms such as DSX to support the deployment of AI computing infrastructure.

The three parties had previously set a long-term construction target of 1 gigawatt of AI computing infrastructure. The current capacity arrangement is centered on the Sejong data center, with gradual expansion overseas in the future.