Inside NAVER, NVIDIA & Brookfield’s Sovereign AI Strategy

Accelerating South Korea’s national AI ambitions, leading tech and AI companies – NAVER, NVIDIA and Brookfield – are coming together in a new proposal that significantly expands the country’s AI factory infrastructure.

Revealed during Korean President Jae Myung Lee’s AI Summit visit to San Francisco, the planned expansion will scale NAVER’s initial multi-tenant NVIDIA DSX AI factory buildout at the GAK Sejong hyperscale data centre in Sejong. 

The original deployment of 55 megawatts will now be tripled, increasing to 200 megawatts by 2028, with NAVER setting a long-term goal of reaching 1 gigawatt of sovereign AI compute.

Additionally, the multi-billion-dollar buildout brings together significant capital commitments across the three companies:

• NVIDIA: Plans to make a direct US$1bn strategic investment into NAVER Corp
• Brookfield: Has entered into a nonbinding term sheet to fund up to US$9bn
• NAVER: Will provide the remaining financial backing to complete the project and operate the facilities.

Strategic expansion of infrastructure

NVIDIA’s planned funding remains subject to standard closing conditions, including NAVER finalising at least US$9bn in committed project financing. 

Once operational, this scale of infrastructure will provide both South Korean and US-based AI innovators with access to production-grade compute needed to build next-generation models, enterprise AI agents and industrial applications.

Haejin Lee, Founder and Chairman of NAVER, says: “NVIDIA’s planned strategic investment and our infrastructure supply agreement with Brookfield have propelled NAVER’s vision for the AI factory business into a robust execution phase. 

“Leveraging the solid partnerships with our global partners, we will drive technological innovation, foster a sovereign AI ecosystem and spearhead efforts to strengthen South Korea’s AI competitiveness.”

Augmenting the core strengths

The partnership brings together the core strengths of each entity to build out a full-stack, lowest-cost sovereign AI factory model for enterprise, industrial and government clients. 

NAVER is contributing deep operational experience as an NVIDIA Cloud Partner with a proven track record of managing hyperscale infrastructure. 

Delivering its industry-leading accelerated computing architecture and comprehensive AI factory platform, NVIDIA will power the facility’s processing needs. 

On the other hand, Brookfield will round out the collaboration by bringing massive capital deployment capabilities and extensive global power infrastructure expertise. 

Jensen Huang, Founder and CEO of NVIDIA, says: “AI factories are the infrastructure nations need to compete and innovate in the intelligence era. 

“Together, NAVER, NVIDIA and Brookfield are building sovereign AI infrastructure at the scale needed to fuel Korea’s startups and industries. This is how nations and companies are building their futures in the age of AI.” 

Brookfield currently manages nearly US$100bn in AI infrastructure value-chain assets, spanning power generation, data centres, compute and semiconductor manufacturing. 

Building on its role as a founding partner in the Brookfield Global AI Infrastructure Program, the latest planned deal marks another milestone in Brookfield’s partnership with NVIDIA.

The global investment firm has an active and long-term presence in the country, already managing approximately US$12bn across real estate, energy and infrastructure.

Sikander Rashid, Global Head of AI infrastructure at Brookfield, says: “As AI adoption accelerates across the global economy, access to trusted, sovereign and scalable AI infrastructure is becoming an increasingly important strategic priority for companies and countries.” 

Advanced hardware and innovation

The expanded 200-megawatt AI factory at GAK Sejong will feature advanced hardware configurations, including the NVIDIA Blackwell and NVIDIA Vera Rubin platforms. 

This is aimed at giving organisations the raw processing power required to deploy complex AI models at scale. 

The buildout relies heavily on the end-to-end NVIDIA DSX platform, which co-designs chips, software, systems and facilities to lower token costs and accelerate production timelines. 

Key components include NVIDIA DSX MaxLPS software to maximise token throughput per megawatt and NVIDIA DSX OS to provide automated health checks, lifecycle management, resilience and multi-tenant management across the facility.

Beyond physical data centre infrastructure, the collaboration deepens NAVER and NVIDIA’s joint work across open models, agentic workflows and physical AI. 

NAVER is advancing its HyperCLOVA X model family using NVIDIA Nemotron 3 Ultra open models, combining them with its proprietary training data and expertise. 

As the first Korean company to join the NVIDIA Nemotron Coalition, NAVER is actively contributing to open model innovation across pretraining, post-training and reinforcement learning. 

Additionally, it plans to launch an enterprise AI agent platform in Korea in the second half of the year using the NVIDIA Agent Toolkit, including NVIDIA NemoClaw blueprints.

Alongside this, the company is developing a digital ‘Seoul World Model’ built on NVIDIA Cosmos world foundation models using proprietary urban street-view and spatial data.