NAVER Signs $1.5 Billion AI Computing Contract With US Customer
NAVER’s $1.5 billion, 20-megawatt US inference deal targets first-half 2027 service as Brookfield-backed assets underpin a 200MW expansion.
By Nathan Brooks
4 min read

What's News
- NAVER disclosed on November 12 a roughly 2 trillion won ($1.5 billion) computing-services contract with an unidentified US AI inference company.
- The deal covers about 20MW of next-generation graphics-processing demand and targets commercial service in the first half of 2027.
- NAVER plans to launch its AI Factory at 55MW in the first half of 2027 and expand to 200MW by 2028.
- Brookfield will invest up to $9 billion through an infrastructure vehicle, while NAVER guarantees related subsidiary obligations for up to eight years.
- Third-quarter enterprise revenue rose 33% to 193.3 billion won, while infrastructure expenses increased 29.1%.
By Nathan Brooks
South Korea's NAVER signed a computing-services contract worth about 2 trillion won ($1.5 billion) with an unidentified US artificial intelligence inference company. NAVER disclosed the deal on its third-quarter earnings call on November 12 and described the company as the initial customer for its AI Factory infrastructure business.
The agreement covers next-generation graphics-processing capacity and represents about 20 megawatts of service demand, according to Yonhap, Newsis and other Korean business publications. NAVER's official third-quarter earnings presentation said customer contracts cover capacity due to come online in “early 2027.”
What does this contract change for NAVER Cloud?
For NAVER Cloud, the answer is narrower than the $1.5 billion headline. The agreement puts 20MW of customer demand behind the AI Factory, but NAVER still has not disclosed the contract period, pricing, service obligations or expected margin.
On NAVER's disclosed figures, 20MW equals roughly 36% of the planned 55MW opening capacity. NAVER expects “commercial service and associated revenue” to begin in the first half of 2027. The signed contract therefore links the first commercial launch to committed customer demand.
NAVER also said it is discussing additional supply agreements with more than 10 prospective customers, including AI companies outside Asia. Those discussions sit outside the current disclosed customer base and could broaden demand beyond the initial 20MW agreement, though they remain supply talks rather than disclosed customer contracts.
Brookfield finances the asset base
NAVER previously said Brookfield would invest up to $9 billion through an infrastructure vehicle. The vehicle would acquire graphics processors and data-center assets for NAVER Cloud to use. NAVER said the new contract builds on its infrastructure agreements with Brookfield and NVIDIA.
Under NAVER's disclosed structure, the Brookfield-backed vehicle would own the computing assets and charge NAVER Cloud usage fees linked to business revenue. NAVER has agreed to guarantee its subsidiary's usage-fee obligations for up to eight years. The company said the guarantee does not itself create an immediate cash outflow.
For NAVER, that structure separates asset ownership from service delivery while retaining a long-dated payment obligation. The practical change is that the $9 billion Brookfield commitment is not a simple upfront capital outlay by NAVER. Investors still have to assess usage fees and revenue together because the guarantee links NAVER's exposure to the economics of the compute assets.
Capacity expands beyond the first contract
NAVER plans to start its AI Factory at 55MW in the first half of 2027 and expand to 200MW by 2028. The 2028 target is about 3.6 times the opening size. NAVER said it has secured data-center capacity beyond the 2028 target, although disclosed customer contracts do not yet cover much of that capacity.
TNGlobal reported in July that NAVER, Brookfield and NVIDIA planned a 200-megawatt Korean AI Factory using NVIDIA computing platforms. It identified NAVER's GAK Sejong data center as the main hub. The 20MW agreement is the first customer marker within that broader program, not the full 200MW buildout.
NAVER leaves key terms undisclosed
NAVER kept the customer unnamed and referred only to “next-generation capacity” without identifying the graphics processors assigned to the agreement. It did not disclose:
- The contract period
- The pricing structure
- Service-level obligations
- Cancellation terms
- Expected margin
The omissions matter because customers cannot assess service reliability, contractual duration or unit economics from NAVER's disclosures alone. The company has provided scale, timing and one customer, but not enough detail to calculate returns from the new business.
Revenue and spending move higher
NAVER reported third-quarter enterprise revenue of 193.3 billion won, up 33% from a year earlier. Infrastructure expenses rose 29.1%, and NAVER attributed part of the higher spending to computing assets and data-center rents. The results show faster enterprise sales alongside heavier infrastructure investment, but NAVER did not disclose the new contract's expected margin or pricing.
NAVER's next proof points are concrete: the company expects commercial service and associated revenue to begin in the first half of 2027 and plans to move capacity from 55MW toward 200MW by 2028. The company is also discussing additional supply agreements with more than 10 prospective customers. Until NAVER discloses additional contracts and the initial deal's economics, the $1.5 billion headline describes contract value, not disclosed profitability.
Source: technode.global
More from Nathan Brooks
Show full bio
News editor covering regulation, policy and their impact on entrepreneurship, startups and venture funding at Business Bearings.
2,046 articles