NVIDIA acquired a 4.5% stake in Naver for 1.48 trillion won, making the GPU maker Naver's third-largest shareholder, the companies announced on the 29th. The investment positions Naver as NVIDIA's strategic AI infrastructure partner in Asia rather than a simple customer relationship, according to securities analysts. Naver is transitioning from an NVIDIA GPU customer to a neocloud operator selling computing resources to external companies, while selecting Brookfield as its exclusive negotiation partner for a minimum $9 billion computing infrastructure project with a total scale of $10 billion by 2028.
NVIDIA Invests Directly in Naver Entity Rather Than Special Purpose Vehicle
NVIDIA's investment targets Naver itself rather than a special purpose vehicle (SPV) established for the AI factory business, according to industry sources on the 29th. Naver issued 7.24 million common shares to NVIDIA through the 1.48 trillion won third-party allocation. The company also plans to cancel 3.1% of treasury shares to reduce shareholder dilution concerns.
The business structure ensures AI factory revenues flow directly to Naver's corporate value rather than being isolated in a joint venture or SPV. "NVIDIA's direct investment in Naver instead of an SPV allows Naver to fully capture AI data center profits," said Lee Jun-ho, researcher at Hana Securities.
Naver's disclosure documents reflect this structure, with NVIDIA's role changed from "business entity jointly bearing sales and business risks" to "GPU supply entity." NVIDIA participates as a shareholder sharing long-term interests, while AI factory operating performance and responsibility belong to Naver.
NVIDIA Stake Increases Naver's Access to Latest GPU Supply
NVIDIA's equity investments in neocloud companies CoreWeave and Nebius provide context for the Naver transaction, with the GPU maker holding approximately 11% and 9.3% stakes respectively. NVIDIA employs this strategy to secure stable sales channels while expanding its software ecosystem through direct investments in GPU demand sources.
Naver's probability of securing latest-generation GPUs increased with the investment, according to analysts. The company targets 55 megawatts (MW) capacity in the first half of 2027, expanding to 100MW by year-end 2027, 200MW in 2028, and 1 gigawatt (GW) domestically and internationally in the mid-to-long term.
"NVIDIA likely selected Naver as a partner whose business areas don't overlap and who possesses easy access to capital essential for data center expansion, while securing software feedback to maintain industry leadership," said Lee Hyo-jin, researcher at Meritz Securities.
CoreWeave represents the most similar business structure to Naver's new venture, having secured GPU priority procurement rights through NVIDIA's equity investment. However, CoreWeave faces rising borrowing costs due to surging debt, while Naver generates over 2 trillion won in annual profit and maintains stable credit ratings from past bond issuances, likely attracting NVIDIA's attention.
![Naver executives and NVIDIA CEO Jensen Huang at AI Summit in San Francisco]()
Naver Chairman Lee Hae-jin and NVIDIA CEO Jensen Huang at the AI Summit in San Francisco on the 24th (local time), where they signed a strategic investment agreement for global AI factory construction.
Naver Combines GPU Infrastructure with Full-Stack AI Capabilities
Naver's competitive advantage extends beyond GPU rental to encompass full-stack capabilities, according to industry analysis. The company accumulated data center construction and continuous operation experience through its 'Gak Chuncheon' and 'Gak Sejong' facilities, while possessing cloud services, proprietary AI models, and sovereign AI services.
Naver can deliver full-stack products combining GPU infrastructure, models, platforms, and operational services based on business experience in public, financial, and manufacturing sectors. The company plans to reduce direct capital burden by utilizing external funding from Brookfield and others in the early business phase, while securing utilization rates through long-term contracts with large anchor customers.
Successful customer acquisition would enable a virtuous cycle where cash from existing advertising and commerce businesses funds AI infrastructure expansion, with AI factory revenues reinvested in growth initiatives. "NVIDIA's partnership enabling neocloud business entry creates expectations for domestic and global revenue, fundamentally changing the company's business structure," said Lee Jun-ho of Hana Securities.
![Naver's Gak Sejong data center facility]()
Naver's 'Gak Sejong' data center
Yoo Seong-man, researcher at Leading Investment Securities, stated: "The collaboration with NVIDIA increases the possibility of stable procurement of latest GPUs while providing a foundation for Naver to offer its accumulated data center construction and operation experience plus cloud and AI model capabilities to global customers."
He assessed the investment as "a core catalyst that will drive Naver's re-evaluation from a domestic internet platform company to an Asian neocloud and global AI infrastructure operator."
FAQ
What stake did NVIDIA acquire in Naver?
NVIDIA acquired approximately 4.5% of Naver through a third-party allocation of 7.24 million common shares worth 1.48 trillion won, making NVIDIA Naver's third-largest shareholder after the transaction completes.
Why did NVIDIA invest directly in Naver instead of a special purpose vehicle?
NVIDIA invested directly in Naver rather than an SPV established for the AI factory business, ensuring that AI data center revenues and business performance flow directly to Naver's corporate value rather than being isolated in a separate entity, according to Hana Securities researcher Lee Jun-ho.
What AI infrastructure capacity does Naver plan to build?
Naver targets 55 megawatts capacity in the first half of 2027, expanding to 100MW by the end of 2027, 200MW in 2028, and 1 gigawatt domestically and internationally in the mid-to-long term, according to the company's disclosed business plans.
