South Korea confirms Nvidia GPU deliveries remain on track

Stable supply for sovereign AI

South Korea is moving forward with its national AI infrastructure goals despite recent headlines surrounding Nvidia's leadership. Vice Minister of Science and ICT Ryu Jemyung confirmed to a parliamentary budget committee that 10,000 Nvidia GPUs funded by a 2025 supplementary budget are already active. These chips are a core component of the country's effort to establish sovereign AI capabilities.

The ministry is actively managing the relationship with the chipmaker to ensure the hardware reaches its destination. Ryu stated that the ministry is closely monitoring developments and is currently in discussions with Nvidia Korea regarding specific prices and delivery schedules. This oversight ensures that the government's investment translates into usable compute power for national projects.

The remarks come at a time when the global AI market is watching how governments secure high-end hardware. By confirming that the 10,000 units are in use, the ministry provides a tangible milestone for its broader deployment plan. This plan aims to scale the country's GPU capacity significantly over the coming years.

Contracted volumes for 2026

Looking ahead, the South Korean government expressed confidence in the long-term pipeline of hardware. Ryu noted that the production and delivery schedules for contracted volumes are expected to proceed without disruption. This is particularly relevant for the 2026 rollout, which remains a priority for the nation's technological roadmap.

The government's commitment is part of a massive build-out announced during the APEC Summit in October 2025. According to the announcement, the Ministry of Science and ICT plans to deploy up to 50,000 GPUs. The 10,000 units currently in use represent about one-fifth of that specific government target.

While the government is a major player, it is not the only entity securing hardware in the region. The broader Korean build-out involves over 260,000 Nvidia GPUs. The ministry's role in this ecosystem is a key piece of the puzzle for South Korea's technological independence.

Corporate and government shares

The distribution of Nvidia's 260,000-plus GPU commitment in Korea shows a mix of public and private investment. The Korean government's share through the Ministry of Science and ICT accounts for about 19% of the total. The remaining 81% is distributed among four major corporate programs involving Samsung Electronics, SK Group, Hyundai Motor Group, and NAVER Cloud.

These companies have significant stakes in the country's AI infrastructure, with each of the four corporations having over 50,000 GPUs in their respective plans. NAVER Cloud specifically has a plan for over 60,000 GPUs. These figures highlight the massive scale of the private sector's involvement in the Korean AI landscape.

The government's 10,000 GPUs in use are part of a plan that started with an initial deployment of 13,000 units. This initial batch was shared among NAVER Cloud, NHN Cloud, and Kakao Corp. The current deployment shows a steady transition from these early partnerships to a broader national infrastructure.

Business continuity in Seoul

Despite internal investigations at Nvidia, the business calendar in South Korea appears to be holding steady. Reports indicate that Nvidia AI Day Seoul, a two-day event for 2,000 participants, is proceeding as planned. The event focuses on infrastructure, agentic AI, and physical AI, which are key areas of interest for the region's tech leaders.

Some minor logistical changes have occurred due to internal circumstances at Nvidia. For example, a planned visit by a senior director to the startup Rebellions was cancelled. However, other partners are maintaining their commitments to the hardware provider.

Kakao Mobility confirmed that its business agreement with Nvidia will proceed as scheduled. While a photo session was cancelled, a senior director will still attend to ensure the partnership remains intact. These actions suggest that while internal corporate matters may cause minor friction, the primary hardware contracts remain robust.

Financing the GPU build-out

The infrastructure expansion in South Korea is supported by unique financial mechanisms. The country recently launched its first GPU-backed loan, which totals 10.7 trillion won. This specific loan is tied to a deal with NAVER Cloud, providing a blueprint for how other entities might finance high-end compute.

This financial framework is essential for meeting the ambitious targets set for 2025 and 2026. By creating dedicated loan structures, the government is attempting to lower the barrier for companies to acquire the massive amounts of hardware required for AI. This helps stabilize the supply chain for domestic tech firms.

As the global compute capex reaches a trillion-dollar scale in 2026, Korea's strategy serves as a model for sovereign AI. The combination of government funding, private corporate investment, and specialized lending creates a multi-layered approach to securing the next generation of computing power.

Next steps for the region

The next phase for South Korea involves scaling the 10,000 units currently in use toward the 50,000-unit goal. The ministry will continue to monitor the delivery of 2026 volumes to ensure they arrive without delay. This will require constant coordination with Nvidia Korea on both logistics and pricing.

Observers will also be watching the progress of the four major corporate programs. Since these companies account for the vast majority of the 260,000 GPUs, their success is a primary indicator of Korea's AI trajectory. Their specific delivery numbers for 2026 have not yet been disclosed.

Finally, the resolution of Nvidia's internal investigation will be a key factor to watch. While the Korean government currently expresses confidence in the supply chain, the outcome of the board's probe could influence future international trade and security discussions regarding high-end chips.