Inside the $10 Billion Financial Engineering Behind Naver's Global Sovereign AI Factory

South Korean internet giant Naver has unveiled the financial blueprints for its ambitious 1-gigawatt (GW) global Naver AI Factory project, demonstrating how next-generation AI infrastructure will actually be funded. By structuring a complex $10 billion partnership with chip giant NVIDIA and alternative asset manager Brookfield, Naver is pioneering a capital-light model that offloads massive hardware balance-sheet risks. The deal marks a shift in how regional cloud players and sovereign AI champions plan to survive the ruinous capital expenditure wars of the generative AI era.

The Multi-Billion Dollar Off-Balance-Sheet Engine

Building gigawatt-scale AI infrastructure is a financial death match that few companies outside of Microsoft, Alphabet, and Meta can afford. Naver’s solution is a masterclass in financial engineering. Instead of buying billions of dollars of GPUs outright—which would devastate its balance sheet and dilute shareholder value—Naver is leveraging an off-balance-sheet Special Purpose Vehicle (SPV) funded primarily by Brookfield.

Under the terms of the Phase 1 agreement, Brookfield will commit up to $9 billion in infrastructure investment directly into the SPV. This entity will own the physical data center assets and the high-end NVIDIA graphics processing units (GPUs). Naver’s operating subsidiary will then pay usage fees to the SPV to access the computing power, subsequently selling these AI resources to its enterprise and sovereign clients. This operational lease structure keeps massive debt off Naver's core balance sheet while securing the infrastructure necessary to compete globally.

"The operational lease structure allows Naver to scale aggressively without incurring the massive debt loads that typically plague capital-intensive infrastructure buildouts."

Ultrathink Analysis

NVIDIA's Sovereign AI Playbook

For NVIDIA, led by CEO Jensen Huang, this deal is about securing long-term distribution and cementation in the sovereign AI sector. NVIDIA is investing $1 billion as direct equity in Naver. This is not passive asset management; it is a strategic maneuver to guarantee that Naver remains locked into NVIDIA's ecosystem as competitive threats from Intel, AMD, and custom hyperscaler ASICs grow.

By taking an equity stake, NVIDIA secures a guaranteed pipeline for its high-margin chips and participates in the downstream revenue generated by the AI Factory. More importantly, it helps NVIDIA build a sovereign AI footprint outside the regulatory reach of the United States and China. Naver’s initial rollout begins at its Gak Sejong data center hub in South Korea, but the blueprint explicitly outlines rapid expansion into Europe and the Middle East—regions desperate to build localized AI capabilities free from American hyperscale hegemony.

Milestones and Global Expansion Scale

The roadmap for Phase 1 is highly aggressive, targeting 200 megawatts (MW) of the eventual 1 GW vision. The timeline is structured to bring capacity online in rapid increments:

• H1 2027: Initial rollout of 55 MW at the Gak Sejong hub.
• End of 2027: Expansion to 100 MW of operational capacity.
• 2028: Achievement of the full 200 MW Phase 1 target.

Once Phase 1 is validated in the Asia-Pacific region, the partnership plans to scale the model internationally. The combination of Brookfield’s deep pockets for real estate and power, NVIDIA’s advanced silicon, and Naver’s operational software stack creates a repeatable, plug-and-play template for sovereign states looking to buy turnkey AI independence.

Why This Matters for the AI Sector

This deal represents a crucial evolution in the business of artificial intelligence. We are moving past the phase of speculative software valuations and entering the era of hard asset financing. The collaboration suggests that the future of large-scale AI computation may look less like traditional software-as-a-service (SaaS) and more like energy grids or telecommunications infrastructure, funded by private equity and secured by long-term capacity contracts.

For founders and enterprise buyers, this model ensures that sovereign AI cloud options will remain viable alternatives to the Big Three US hyperscalers. For investors, it highlights a new asset class: GPU-backed infrastructure funds where institutional capital can seek yield in the AI buildout without taking direct equity risk in volatile AI software startups.

The Sovereign Path Forward

The Naver AI Factory blueprint proves that regional tech giants do not have to capitulate to Silicon Valley's capital dominance. By splitting the burden—allowing Brookfield to take the asset risk, NVIDIA to take the technology risk, and Naver to take the operational and market risk—the trio has built a model that could define the next decade of global computing infrastructure. The sovereign AI race is no longer just about who has the best algorithms; it is about who has the smartest bankers.