Nvidia’s latest Korea announcements strengthen the long-term case for its AI infrastructure business, but investors should distinguish proposed commitments from completed transactions. Nvidia plans to invest $1 billion in NAVER Corp. as part of a $10 billion sovereign-AI infrastructure project, while its expanded partnership with SK Group carries a stated value of more than $500 billion. The NAVER investment remains conditional, and the SK framework is currently supported by letters of intent rather than a disclosed definitive agreement. The scale is still notable.
NAVER intends to expand the initial deployment at its GAK Sejong data center from 55 megawatts to 200 megawatts by 2028, using Nvidia’s Vera Rubin and Blackwell platforms. Separately, SK Telecom plans an Nvidia Vera Rubin and DSX AI factory with capacity of up to 2 gigawatts in Korea, with the first facility targeted to come online in 2027. These projects illustrate how Nvidia is positioning its computing platforms at the center of national-scale AI development. For Nvidia stock, the announcements offer evidence of sustained demand across GPUs, networking, systems and software. They do not, however, establish when revenue will be recognized, how much of the announced project value will flow to Nvidia, or whether every planned phase will proceed on schedule.
Table of Contents
- What Do the $1 Billion NAVER Investment and Expanded SK Group Partnership Mean for Nvidia Stock?
- How the NAVER Sovereign-AI Project Is Structured
- SK Group Expands Nvidia’s Role in AI Factories and Memory
- What Nvidia Investors Should Monitor Next
- Valuation Risks and Common Interpretation Errors
- Korea’s Importance in Nvidia’s AI Supply Chain
- The Capacity and Contract Milestones Behind the Headlines
- Frequently Asked Questions
What Do the $1 Billion NAVER Investment and Expanded SK Group Partnership Mean for Nvidia Stock?
The NAVER agreement could deepen Nvidia’s role beyond that of a hardware supplier. Nvidia says it plans to invest $1 billion directly in NAVER, subject to customary closing conditions and NAVER obtaining at least $9 billion in committed project financing. Brookfield has entered a nonbinding term sheet to provide up to $9 billion, while NAVER would fund the remaining amount required for the stated $10 billion project. That structure matters to shareholders because an equity investment and a product sale have different financial effects. A server order can contribute to revenue when products are delivered and accounting requirements are met.
An investment in NAVER would instead place Nvidia’s capital at risk in exchange for strategic and potential financial benefits. For comparison, Nvidia could benefit as a supplier to the project even if the value of its NAVER stake later fluctuates. The SK Group announcement is much larger in headline terms, describing plans for a comprehensive partnership valued at more than $500 billion. That figure should not be treated as an Nvidia order backlog or expected Nvidia revenue. The planned collaboration covers AI-factory construction and AI-memory supply across multiple SK businesses, and the parties signed letters of intent to formalize the arrangement.
How the NAVER Sovereign-AI Project Is Structured
NAVER, Nvidia and Brookfield announced the proposed expansion of NAVER’s Korean sovereign-AI infrastructure on july 24, 2026. The project would increase the initial DSX deployment at the GAK Sejong data center from 55 megawatts to 200 megawatts by 2028. NAVER also states a longer-term ambition of reaching 1 gigawatt, although that larger figure is not current installed capacity or a guaranteed construction commitment. The planned 200-megawatt AI factory is expected to use Nvidia’s Vera Rubin and Blackwell platforms and serve AI developers in Korea and the United States.
This could create demand extending beyond accelerators to networking, rack-scale systems and Nvidia’s software ecosystem. A real-world comparison is a cloud data center that standardizes on one computing architecture: the initial hardware deployment can encourage customers to build applications around the same platform, making later expansion easier. investors should not describe the project as fully financed. Brookfield’s agreement is a nonbinding term sheet for up to $9 billion, and Nvidia’s planned $1 billion investment depends partly on NAVER securing at least $9 billion in committed financing. Financing costs, power availability, construction delays and changes in AI demand could affect the project’s eventual size and schedule.
SK Group Expands Nvidia’s Role in AI Factories and Memory
The proposed SK partnership combines two strategically connected parts of the AI supply chain. SK Telecom would build an Nvidia Vera Rubin and DSX AI factory with capacity of up to 2 gigawatts in Korea, while Nvidia and SK hynix would enter a long-term AI-memory partnership intended to secure supply and jointly develop and optimize next-generation memory, including high-bandwidth memory. HBM is particularly important because advanced AI accelerators need rapid access to large volumes of data. A powerful processor can be constrained if memory cannot supply data quickly enough, much as a factory’s output can be limited by a slow component-delivery system.
Closer coordination with SK hynix could help Nvidia align future accelerator designs with memory capacity, performance and power requirements. The announcement extends an existing relationship rather than creating one from scratch. In October 2025, Nvidia said SK Group was building an AI factory with more than 50,000 Nvidia GPUs, with the first phase scheduled for completion in late 2027. That earlier arrangement also included cooperation on HBM and advanced memory, providing a concrete foundation for the broader july 2026 plans.
What Nvidia Investors Should Monitor Next
Investors should track milestones instead of relying only on the announced dollar figures. For NAVER, the most immediate checkpoints are committed project financing, satisfaction of closing conditions and confirmation that Nvidia’s $1 billion investment has closed. Later indicators include construction progress at GAK Sejong, equipment deployment and movement from 55 megawatts toward the planned 200-megawatt capacity. For SK Group, useful signals include the conversion of letters of intent into definitive contracts, disclosed purchase commitments, construction starts and evidence that the first AI factory remains on track for 2027.
Investors should also watch Nvidia’s regulatory filings and earnings commentary for information about purchase obligations, customer concentration, expected shipments and revenue timing. There is a tradeoff between strategic reach and capital discipline. Investing directly in NAVER could reinforce demand for Nvidia’s platform and help establish a major sovereign-AI customer, but it also commits Nvidia capital that could otherwise support research, acquisitions or shareholder returns. Selling equipment without investing would reduce direct capital exposure, while potentially giving Nvidia less influence over the project’s architecture and expansion.
Valuation Risks and Common Interpretation Errors
A common mistake is to add the $10 billion NAVER project and the $500-billion-plus SK partnership directly to Nvidia’s future revenue. Neither headline figure represents a disclosed Nvidia sales contract. The NAVER total includes infrastructure financing and NAVER’s own contribution, while the SK figure covers a broad partnership involving AI factories and memory supply across multiple companies and years. Timing is another limitation.
The NAVER expansion targets 200 megawatts by 2028, and SK Telecom’s first factory is planned for 2027. Even if equipment orders are finalized, revenue could arrive in stages and may depend on construction, electrical capacity, permits, supply availability and customer acceptance. Delays could shift recognized revenue without necessarily canceling the underlying project. Nvidia also expressly identifies the partnerships, expected benefits, product availability and technology development as forward-looking matters subject to risks and uncertainties. Vera Rubin systems must reach customers at the expected performance, cost and production scale, while Blackwell deployments must proceed alongside the industry’s continuing power and cooling constraints.
Korea’s Importance in Nvidia’s AI Supply Chain
Korea offers Nvidia both large infrastructure customers and critical memory expertise. NAVER provides cloud, internet and AI services; SK Telecom contributes telecommunications and data-center capabilities; and SK hynix supplies HBM used in advanced AI systems.
Bringing these roles together can support a more integrated path from memory development to AI-factory operation. The SK plan provides a specific example of that integration: Nvidia and SK hynix intend to secure AI-memory supply while jointly developing and optimizing future memory, and SK Telecom plans to deploy Nvidia’s Vera Rubin and DSX architecture in an AI factory of up to 2 gigawatts.
The Capacity and Contract Milestones Behind the Headlines
The disclosed milestones give investors a practical framework for separating current operations from planned scale. NAVER’s project starts with a 55-megawatt DSX deployment, targets 200 megawatts by 2028 and states a longer-term ambition of 1 gigawatt.
SK Telecom’s proposed AI-factory network could reach up to 2 gigawatts, with the first factory planned for 2027. Contract status remains equally important: Brookfield has signed a nonbinding term sheet for up to $9 billion, Nvidia’s planned $1 billion NAVER investment remains subject to closing conditions and a financing threshold, and the expanded SK partnership is presently based on letters of intent.
Frequently Asked Questions
Has Nvidia completed its $1 billion investment in NAVER?
No. Nvidia plans to invest $1 billion, but the transaction remains subject to customary closing conditions and NAVER securing at least $9 billion in committed project financing.
Is Brookfield’s $9 billion commitment binding?
The announced arrangement is a nonbinding term sheet under which Brookfield could provide up to $9 billion. It should not be treated as completed financing.
Does the $500-billion-plus SK partnership represent Nvidia revenue?
No. It describes the stated scale of a broad partnership spanning AI-factory construction and AI-memory supply. Nvidia has not disclosed that amount as revenue, backlog or a firm product order.
When are the Korean AI factories expected to become operational?
SK Telecom’s first planned factory is targeted to come online in 2027. NAVER aims to expand its GAK Sejong deployment from 55 megawatts to 200 megawatts by 2028.
Which Nvidia platforms are involved?
The planned NAVER facility is expected to use Blackwell and Vera Rubin platforms. SK Telecom’s proposed AI factory would use Nvidia Vera Rubin and DSX infrastructure.
Why is SK hynix important to Nvidia?
SK hynix produces high-bandwidth memory, a critical component for AI accelerators. The companies intend to secure supply and jointly develop and optimize next-generation AI memory.