Analysis of the NVIDIA-Naver Global AI Infrastructure Alliance and Nationwide Strategic Cooperation in South Korea
Jensen Huang, CEO of NVIDIA, a leader in the global AI semiconductor market, recently visited South Korea for a series of high-level meetings with major IT, consumer electronics, automotive, and semiconductor conglomerates. This visit moved beyond simple supply chain coordination, focusing on establishing a strategic alliance to co-construct the next-generation AI ecosystem—specifically “AI Factories” and “Physical AI” infrastructure—using Korea as a central hub.
The proposal for a Gigawatt (GW)-scale Global AI Factory presented to NAVER is particularly significant. It marks a major milestone for Naver, shifting its business model from a domestic consumer platform to a global B2B AI infrastructure provider. This report analyzes the details of the proposal, the ripple effects on the Korean industry, and the core focus areas for investors.
1. The Prolouge: Jensen Huang’s Visit to Naver ‘1784’
Jensen Huang visited Naver’s second headquarters, “1784,” in Bundang, Gyeonggi Province, to meet with Lee Hae-jin, Naver’s Global Investment Officer (GIO). The 1784 building serves as a testbed for advanced technologies such as digital twins, robotics, and 5G private networks. Huang praised Naver as a “company of the future” after being served iced coffee by an autonomous robot.
The camaraderie between the two leaders was evident in a comic strip event where Huang wrote “Don’t worry! I have GPU!” in response to Lee’s note about work and happiness. This followed a previous “Samgyeopsal (Pork Belly) and Soju” meeting between the heads of SK, LG, and Naver, signaling a united front between NVIDIA and Korea’s hardware and software giants.
2. The 1GW Global AI Factory Roadmap
The core of the alliance is the construction of an “AI Factory”—a massive intelligent facility optimized for the “production of tokens at the lowest cost and highest speed.” By combining Naver’s hyperscale data center expertise with NVIDIA’s “DSX” high-performance computing platform, the goal is to achieve 1GW of capacity.
| Phase | Timeline | Cumulative Capacity | Strategy & Partnership Details |
|---|---|---|---|
| Phase 1 | H1 2027 – End of 2028 | 200MW | Secure capacity in Korea, Malaysia, and Japan via leasing. Target: 55MW in early 2027, reaching 200MW by late 2028. |
| Phase 2 | 2029 – 2030 | 400MW – 500MW | Expansion linked to Naver’s ‘GAK Sejong’ data center and proprietary hyperscale design. |
| Phase 3 | 2029 – 2030 | 600MW – 800MW | Additional leasing contracts in domestic and global strategic hubs. |
| Phase 4 | Post-2030 | 1GW (1,000MW) | Greenfield construction of eco-friendly data centers targeting Asia, the Middle East, and Europe. |
To reach the initial 200MW, Naver and a strategic partner (likely NVIDIA or a major anchor client) will establish a Special Purpose Vehicle (SPV), with each side contributing approximately $1 billion. The total cumulative investment required for 1GW is estimated at $50–$60 billion, to be secured through long-term project financing.
Core Technology Pillars
- Nemotron Alliance: Naver is the first and only Korean company to join the NVIDIA-led “Nemotron Alliance” (alongside 12 global leaders like Mistral AI). Naver will use “Nemotron 3 Ultra” technology to enhance the performance and global versatility of its “HyperCLOVA X” model.
- Cloud Standardization (DSX): Naver will integrate its GPU cluster management knowledge with NVIDIA’s Unified High-Performance Computing architecture (DSX), standardizing everything from design to deployment.
- Physical AI (Seoul World Model): Naver’s 3D spatial modeling data will be integrated into NVIDIA’s “COSMOS” platform to create the “Seoul World Model,” a virtual simulator for autonomous robotics and smart city management.
3. Industrial Matrix: Major Group Alliances
Jensen Huang emphasized that Korea is unique in possessing world-class capabilities across manufacturing, electronics, and software.
- Hyundai Motor Group (Physical AI Hub): Chairman Euisun Chung proposed a $9 billion “Saemangeum AI Valley” project in Gunsan, focused on robotics, hydrogen energy, and autonomous driving. NVIDIA promised to invest in data centers there, with Huang noting that the “Silicon Valley of Physical AI is being invented in Saemangeum.”
- SK Group (Memory Co-Design): Chairman Chey Tae-won confirmed SK hynix as NVIDIA’s “most powerful memory ally.” They agreed to “co-design” high-bandwidth memory (HBM) for next-gen products like the “Vera Rubin” AI accelerator and “Jetson Thor” robotics boards.
- LG Group (Thermal Management & Robotics): LG will provide precision cooling solutions (chillers) and liquid cooling for high-heat Blackwell/Rubin GPUs. Additionally, LG’s robotics platforms (Bear Robotics, Robostar) will integrate NVIDIA’s Isaac and GR00T platforms.
- Samsung Electronics (Supply Resilience): Discussions focused on HBM4E/HBM5 quality tests and long-term foundry outsourcing for automotive chips. Samsung’s diversified portfolio (Apple, Amazon, Google) ensures it remains a stable, non-dependent partner.
4. Market Outlook & Expert Consensus
Analysts suggest that Naver is undergoing a “valuation re-rating,” shifting from a domestic portal to a global B2B infrastructure provider.
| Brokerage | Target Price (KRW) | Key Rationale |
|---|---|---|
| DS Investment | 450,000 | The AI Factory adds at least 19 trillion KRW in enterprise value. |
| Hana Securities | 400,000 | Evolution into the “CoreWeave of Asia” justifies a higher valuation multiple. |
| Kyobo Securities | 390,000 | Addition of a 20 trillion KRW annual revenue stream from global infrastructure. |
The “Sovereign AI” Premium
Experts highlight the “Sovereign AI” model as a powerful export. Nations wary of US-China tech dominance are looking for a “Third Way”—localized AI stacks that respect data sovereignty. This provides Naver with a recurring revenue model previously unseen in the Korean tech sector.
5. Investor Blueprint: Key Companies to Watch
Investors should focus on the transition from “component suppliers” to “system-level partners.”
| Company | Momentum | Critical Risk |
|---|---|---|
| NAVER | Global CSP expansion via 1GW AI Factory; Nemotron Alliance membership. | High CAPEX and interest costs during the construction phase. |
| LG Electronics | Global surge in demand for data center cooling (Chillers) and industrial robots. | Volatility in the global appliance cycle; robot revenue scaling speed. |
| Hyundai Motor | Physical AI implementation in smart factories and autonomous driving hubs. | EV market slowdown; regulatory delays in hydrogen infrastructure. |
| SK hynix | Exclusive design partner for NVIDIA’s next-gen HBM architectures. | Massive costs for new HBM testing fabs; competition in standards. |
| Samsung Electronics | Potential for major foundry wins and HBM5 qualification. | Macro-cycle sensitivity of commodity memory; foundry lead times. |
Conclusion for Investors
The NVIDIA-Korea alliance proves that Korean tech companies have evolved into the “forward operating bases” for future physical infrastructure, encompassing smart manufacturing, logistics, and green energy grids. Investors should monitor long-term contract signings and Special Purpose Vehicle (SPV) equity structures rather than chasing short-term market hype.