[Edaily Reporter Lee So-Hyun ] The details of the first-phase contract structure for the global artificial intelligence (AI) factory project being pursued by Naver (NAVER(035420)), NVIDIA, and Brookfield have come into focus.

The detailed project structure for the gigawatt (GW)-class AI infrastructure plan, first unveiled last June, has now been revealed. Naver plans to secure a stable supply of graphics processing units (GPUs) through its equity partnership with NVIDIA and reduce the financial burden associated with large-scale facility investments by leveraging Brookfield’s investment structure.

On the 3rd, Naver announced “Naver AI Factory Phase 1” through a regulatory filing.

The plan includes a $1 billion equity investment from NVIDIA and an infrastructure investment of up to $9 billion from Brookfield. This Phase 1 project targets a 200 MW infrastructure component of the total 1 GW AI Factory construction plan that Naver announced last June.

The core of the structure is that NVIDIA will make a $1 billion equity investment in Naver, while Brookfield will raise the funds necessary to secure GPUs and data center facilities through a special purpose vehicle (SPV). Naver will oversee its AI computing business through a wholly-owned subsidiary that operates the AI factory. Under this structure, the operating subsidiary will procure computing resources from the SPV, pay for them based on usage, and generate revenue by selling the secured resources to external customers such as corporations and institutions.

Securing GPUs and Distributing Financial Burden Are Key

This structure is significant in that it demonstrates Naver is not pursuing its AI infrastructure business solely through large-scale direct capital investment. As competition in the AI data center market intensifies and securing cutting-edge GPUs has emerged as a key challenge for global companies, this move is seen as a strategic step to stabilize the GPU supply base by establishing an equity relationship with NVIDIA.

If NVIDIA participates not merely as a supplier but as a Naver shareholder, this could serve as a safeguard in negotiations over GPU supply terms and pricing. From NVIDIA’s perspective, expanding its GPU ecosystem through Naver’s growth in the AI computing business aligns with its interests, creating a structure where the interests of both companies converge.

The special purpose vehicle (SPV) being established by Brookfield is a mechanism to reduce the burden of large-scale investment. Discussions are currently underway under a model where, instead of Naver directly purchasing both GPUs and data center equipment, the SPV would own the equipment and Naver’s operating entity would pay usage fees. Through this arrangement, Naver can secure the resources necessary for its AI computing business while reducing the burden of initial investment and direct borrowing.

However, the specific accounting treatment may vary depending on the final contract terms, control over the assets, and usage obligations. Naver is also considering acquiring a minority stake in the SPV to ensure a stable supply of computing resources.

First Hub for ‘Gak Sejong’… Targeting 200 MW by 2028

Naver previously announced on June 8 that it would pursue a joint project with NVIDIA to build a global AI factory. At the time, Naver revealed its plan to jointly build and operate a large-scale AI factory with NVIDIA in response to the surging demand for global AI infrastructure, with the aim of securing new revenue streams in the global AI compute market.

According to the announcement, Naver will lead the acquisition of data center sites as well as construction and operations, while NVIDIA will participate as a business partner responsible for supplying GPUs, acquiring global customers, and sharing revenue and business risks.

Naver has set a tentative goal to expand its infrastructure, starting with a 55 MW capacity in the first half of 2027, reaching a cumulative 100 MW by the end of 2027, and a cumulative 200 MW by 2028. Ultimately, the company plans to build GW-scale infrastructure.

The first hub will be the hyperscale data center “Gak Sejong.” Starting with Gak Sejong, Naver plans to gradually expand its business footprint beyond the Asia-Pacific region to Europe, the Middle East, and other global markets. From a technical perspective, the company will adopt NVIDIA’s next-generation infrastructure platform, “DSX,” to streamline data center operations and enhance business viability.

The key challenges are securing external customers and finalizing specific contracts. Naver plans to generate revenue by selling computing resources secured by the AI Factory operator to companies and institutions. Brookfield has been granted exclusive negotiation rights by Naver for 12 weeks. If negotiations are concluded within this period, Brookfield will be selected as the final financing partner; if negotiations fall through, Naver may resume discussions with other investors.

Naver plans to disclose the progress of this project and the status of major contract signings at each stage by December 31. The company has stated that it will immediately disclose any significant developments, such as contract signings, even before that date. Potential obstacles cited include delays in securing local data center sites, obtaining permits in various countries, securing power infrastructure, and fluctuations in the global supply of AI infrastructure.