Nvidia announces $10 billion AI infrastructure expansion in South Korea, including $1 billion investment in Naver partnership

Nvidia Corp. has announced a $1 billion investment in South Korean internet giant Naver Corp. to finance the major expansion of an AI data center under construction in the country. The investment is part of a broader $10 billion financing package that will enable Naver to more than triple the capacity of its GAK Sejong hyperscale data center in Sejong, South Korea, expanding it from 55 megawatts to 200 megawatts by 2028.

Under the agreement, Canadian asset manager Brookfield will serve as the project's exclusive capital partner, funding up to $9 billion through a nonbinding term sheet. Nvidia contributes $1 billion, with Naver covering the remaining financing needed to complete the project.

The expanded facility will run on Nvidia's AI computing hardware, including its Vera Rubin and Blackwell chip platforms, designed to give both Korean and U.S.-based AI developers access to production-scale computing power for building next-generation AI models, agents, and services. The facility will leverage Nvidia's DSX AI factory platform. Naver has indicated plans to eventually expand its deployment of Nvidia infrastructure to a full gigawatt of capacity.

Naver founder and chairman Haejin Lee credited the new financing with accelerating the company's AI ambitions, stating that Nvidia's investment and the infrastructure agreement with Brookfield had "propelled NAVER's vision for the AI factory business into a robust execution phase." Naver will continue developing its HyperCLOVA X AI models using Nvidia's Nemotron open-source models and has joined the Nemotron Coalition, a group focused on open AI model development. Nvidia's investment remains subject to standard closing conditions, including Naver securing at least $9 billion in committed financing separate from Nvidia's contribution.

Nvidia simultaneously announced a separate commercial partnership with South Korea's SK Group, valued at more than $500 billion in total business over time. This figure includes memory chip purchases from SK Hynix, currently the world's largest supplier of high-bandwidth memory used in AI systems, as well as SK Group's purchases of Nvidia's AI supercomputers. Nvidia will also work directly with SK Hynix to help design future generations of high-bandwidth memory chips to secure reliable access to this critical component amid global AI data center buildout.

SK Telecom plans to build more than 2 gigawatts of AI data centers across the Korean Peninsula—sufficient power to supply approximately 1.5 million homes. The first of these AI factories is expected to open next year.

Nvidia Chief Executive Jensen Huang emphasized the partnership's scale in an interview with Bloomberg Television, stating: "So between us, we're going to do half a trillion dollars' worth of business." Huang also highlighted the broader significance of Nvidia's expanding presence in South Korea, describing the current moment as a turning point for the country's technology sector. "This is the golden ages for Korea," he said, pointing to the country's semiconductor manufacturing strength and industrial base as key reasons for Nvidia's continued investment.

Both announcements coincided with a visit to Silicon Valley by South Korean President Lee Jae Myung, who traveled to San Francisco for an AI summit where the deals were formally unveiled. Additional technology agreements between Korean and American companies are reportedly still being finalized as part of the broader initiative tied to President Lee's visit.

These Korea-focused deals extend Nvidia's aggressive dealmaking strategy over the past year to secure chip supply chains and downstream infrastructure supporting global AI computing demand. Nvidia's relationship with South Korean technology companies has deepened in recent months following earlier chip supply agreements with Samsung, Hyundai, and SK Group during a prior visit by Huang to the country.

South Korea has separately set a goal of deploying approximately 200,000 high-performance GPUs by 2030 as part of its national AI strategy, though the country continues to face infrastructure challenges related to energy supply and data center cooling capacity as it scales its AI computing footprint. With Naver working to finalize the remaining $9 billion in project financing, the full Sejong data center expansion is not expected to be completed until 2028. The scope of the announcements—spanning direct equity investment, chip supply agreements, and multibillion-dollar infrastructure commitments—underscores South Korea's centrality to Nvidia's strategy for securing both chip manufacturing capacity and physical infrastructure to sustain global AI demand.