Korea's Shinsegae, US-based Reflection AI to build Korea's largest AI data center
Shinsegae Group and Reflection AI of the United States announced a strategic partnership to build what they describe as Korea’s largest AI data center, with an initial target capacity of 250 megawatts. The plan calls for a staged expansion and the establishment of a joint venture this year, following the signing of a memorandum of understanding on the evening of the 16th at the National AI Center in San Francisco.
The signing event was attended by Shinsegae Group Chairman Jung Yong-jin and Reflection AI CEO Misha Laskin, who pledged to advance their collaboration. The two firms said they would move quickly to coordinate with relevant government agencies and local authorities as they proceed.
Under the agreement, the partners plan to develop a full-stack AI factory built around a large data center. The goal is to provide cloud services alongside tailored AI solutions, with an emphasis on offering an AI cloud service that both Korean government agencies and private sector companies can use. This aligns with Korea’s push to strengthen AI competitiveness and build sovereign AI capabilities.

In remarks accompanying the signing, Shinsegae executives described the project as foundational for the group’s future growth and for elevating Korea’s overall AI ecosystem. They said the initiative is intended to support a broader shift toward AI as a future growth engine across the conglomerate and the domestic economy.
A notable public endorsement came from a U.S. official who attended the event, underscoring Washington’s support for the project. The presence of the U.S. Commerce Department figure highlighted a broader context: the partnership is framed as the first technology cooperation under the U.S. government’s AI export program begun last July.
Reflection AI’s leadership emphasized Korea’s status as a global IT powerhouse and stressed a joint effort to develop AI infrastructure that Korea can lead. The two companies also announced that the data center would source its GPUs from Nvidia, a supplier that has attracted significant investment from Reflection AI and other partners in recent years.

Nvidia’s participation is significant because GPUs are central to modern AI training and deployment, and the companies’ expectation of reliable GPU supply is tied to recent capital infusions—Reflecting Nvidia’s reported $2 billion investment in AI hardware partnerships last year. The arrangement suggests a strategic overlap of global semiconductor and AI cloud ecosystems with implications for regional and cross-border supply chains.
For U.S. readers, the deal matters for several reasons. It signals deeper cooperation between a major Korean conglomerate and a U.S. AI firm on building advanced AI infrastructure in Asia, with potential implications for cloud markets, data localization, and cross-border technology transfer. It also ties into U.S. policy interests around AI competitiveness, export controls, and the secure, resilient deployment of AI technologies that involve American hardware suppliers and software ecosystems.
Korean authorities and businesses see the project as a potential accelerator of the country’s sovereign AI ambitions, combining domestic cloud services with international investment and technology partnerships. If realized, the data center and its full-stack AI factory could influence regional AI capacity, cloud pricing, and the availability of government-facing AI services in Korea, while contributing to broader U.S.-Korea cooperation in AI research, security, and digital infrastructure.