Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Naver in talks for Asia's largest Nvidia server order, signs 2 trillion won AI deal

Naver said on its third-quarter earnings call that it is negotiating what would be Asia's largest server purchase order with Nvidia to secure Vera Rubin GPUs in advance. The company has also signed a roughly 2 trillion won computing services contract with a U.S. AI inference firm and expects its AI factory business to post operating profit from 2027.

Published

Naver is turning AI infrastructure into an external business. On its third-quarter earnings call on Nov 12, the company said it is negotiating what would be Asia's largest server purchase order with Nvidia, aiming to lock in next-generation Vera Rubin GPUs ahead of demand.

CEO Choi Soo-yeon told the call that the data center capacity Naver has already contracted and completed exceeds the target volume it had disclosed through 2028.

That capacity belongs to a global AI factory project with Nvidia and Brookfield. The plan is to launch services with 55 MW in the first half of 2027, scale to 100 MW by the end of that year and reach 200 MW in 2028, with a long-term goal of 1 GW of AI computing infrastructure.

The first revenue is already booked. Naver signed a computing services contract worth about 2 trillion won with a U.S. AI inference company. CFO Kim Hee-cheul said the deal covers roughly 20 MW and will generate about 2 trillion won over five years, using a mix of B300 and Vera Rubin GPUs on four- to five-year terms, with initial volumes delivered in stages from 2027.

On profitability, Naver expects the AI factory business to post operating profit starting in 2027. Kim cautioned that a two- to three-month gap between equipment arrival and service start will keep margins low in the first year, but said margins should reach well into the double digits once operations run at full capacity.

The customer pipeline is widening. Naver says it is discussing collaboration contracts with at least 10 companies, including counterparties in North America and outside Asia, and will announce signings in sequence.

Naver argues its edge is not only access to GPUs. Running chips like Vera Rubin at scale requires high-density power and liquid cooling, and few data centers are equipped for it. Choi said the market is marked less by oversupply than by a shortage of the latest GPUs, especially Vera Rubin, and of liquid-cooled space to host them.

Its own AI services are moving too. After AI Briefing and AI Tab, Naver plans to launch a personal AI agent app within the year, connecting data from search, shopping, Place, reservations, Naver Pay and user-generated content to understand intent and take real actions. It will also set up a logistics-focused subsidiary.

For Naver, the contract is proof of demand and initial business viability. The real test comes in 2027: whether deliveries land on schedule, whether liquid-cooled capacity keeps pace, and whether the ten-plus negotiations convert into signed deals.

Why it matters

The contract moves Naver from search company to regional AI capacity supplier, and shows that in a tight GPU cycle liquid-cooled, high-density data center space is itself the scarce asset. Delivery timing in 2027 and the ten-plus negotiations will test whether the model repeats.

NaverNvidiaAI Infrastructure
Back to realtime news

Nearby Updates

All