NAVER, NVIDIA, and Brookfield plan to triple Korea's flagship sovereign AI factory to 200 megawatts

NAVER, NVIDIA and Brookfield announced a plan to more than triple the capacity of Korea's flagship sovereign AI data center, with NVIDIA committing $1 billion and Brookfield offering up to $9 billion in project financing. The expansion grows the NVIDIA DSX AI factory at NAVER's GAK Sejong data center from 55 megawatts to 200 megawatts by 2028, with NAVER laying out a longer-term goal of reaching 1 gigawatt of NVIDIA AI infrastructure.

What's new

NAVER, NVIDIA and Brookfield "today announced a proposed expansion of Korea's sovereign AI factory infrastructure, with planned investments that will grow the initial NVIDIA DSX AI factory deployment to 200 megawatts — more than tripling the 55-megawatt buildout announced last month," according to NVIDIA's newsroom announcement. NAVER separately "intends to expand its deployment of NVIDIA AI infrastructure to 1 gigawatt."

The financing structure splits three ways: NVIDIA plans to invest $1 billion directly into NAVER Corp, Brookfield has signed a nonbinding term sheet to fund up to $9 billion, and NAVER will cover the remainder. NVIDIA's investment is contingent on customary closing conditions and on NAVER first lining up at least $9 billion in committed financing for the project — a threshold Brookfield's term sheet is designed to clear.

The buildout runs on NVIDIA's DSX platform, the company's reference architecture for gigawatt-scale "AI factories" — purpose-built data centers optimized for training and running large models rather than general-purpose cloud workloads. GAK Sejong is NAVER's existing hyperscale data center campus and the anchor site for the expansion.

Executives from all three companies framed the deal around sovereignty and national AI capacity. NAVER founder and chairman Haejin Lee said "leveraging the solid partnerships with our global partners, we will drive technological innovation, foster a sovereign AI ecosystem and strengthen South Korea's AI competitiveness." Brookfield's global head of AI infrastructure, Sikander Rashid, said "access to trusted, sovereign and scalable AI infrastructure is becoming an increasingly important strategic priority for companies and countries." NVIDIA founder and CEO Jensen Huang added that "AI factories are the infrastructure nations need to compete and innovate in the intelligence era."

Context

This is the second major Korea-focused AI infrastructure announcement from NVIDIA in as many days. On July 24, NVIDIA and SK Group unveiled a separate $500 billion-plus partnership spanning a 2-gigawatt SK Telecom AI factory and a long-term AI memory supply deal with SK hynix. Together, the NAVER/Brookfield expansion and the SK Group partnership represent two parallel tracks of South Korea's push to build sovereign AI compute capacity domestically rather than renting it from US hyperscalers — a strategy the government has been actively courting, alongside NVIDIA's earlier joint research lab with KAIST announced July 23.

NAVER, which operates South Korea's dominant search engine and a wide portfolio of consumer and cloud services, has positioned itself as the country's leading sovereign-AI operator, distinct from the telecom-led SK Group effort. The 55-megawatt GAK Sejong deployment referenced in this release was itself only announced last month, making this expansion a rapid scale-up rather than a new project from scratch.

Why it matters

The deal is a concrete data point in the broader "sovereign AI" trend, where national governments and domestic champions build and control their own AI compute rather than depending entirely on foreign cloud providers. For NVIDIA, locking in multiple, competing domestic buyers in the same country — NAVER and SK Group both building gigawatt-scale DSX deployments — diversifies its Korean revenue base and reduces reliance on any single customer relationship.

The financing structure is also notable: NVIDIA taking an equity-like stake in a customer (NAVER) while a financial investor (Brookfield) backstops the debt is a pattern NVIDIA has repeated elsewhere as it works to accelerate AI factory buildouts globally. If the 1-gigawatt long-term target materializes, NAVER would become one of the larger single-country AI compute operators outside the US and China, underscoring how quickly "AI factory" scale has shifted from hundreds of megawatts to gigawatt-class infrastructure within a single national market.