Particle.news

Nvidia Signs Multi‑Year Korean Pacts to Build AI Factories and Co‑Develop Memory

The agreements seek to lock in advanced memory supply and to deploy onshore, gigawatt‑scale AI data centers that underpin robotics and physical AI development.

Overview

  • Nvidia formalized a package of deals with South Korean firms on Monday that includes a multi‑year memory co‑development partnership with SK hynix to support next‑generation HBM for Nvidia’s Vera Rubin systems.
  • SK Telecom and Nvidia committed to build a gigawatt‑scale AI cloud in Korea with the first purpose‑built ‘AI factory’ targeted to begin operations in 2027.
  • Naver pledged to expand its Sejong data center using Nvidia’s DSX platform starting at 55 megawatts in early 2027, with a stated plan to scale to gigawatt levels over time.
  • Separately, Nvidia announced technology collaborations with LG on humanoid robotics and future data‑center architecture, with Hyundai and Doosan signing on for work in physical AI, mobility and robotics energy solutions.
  • Executives warned the global shortage of advanced memory will last years, so the partnerships emphasize multi‑year road‑mapping, capacity planning and close technical coordination rather than one‑off contracts.