Anthropic opens Seoul office and signs government MOU as NAVER, Samsung SDS, and LG CNS deploy Claude
Anthropic opened a permanent office in Seoul on June 17, 2026. The same day, the company announced an MOU with Korea's Ministry of Science and ICT and disclosed Claude deployments at five of the country's largest technology organizations.
What happened
KiYoung Choi leads the Seoul office as Representative Director of Anthropic Korea, a role drawing on three decades of technology leadership in the Korean market.
The government MOU commits Anthropic to working with the Ministry of Science and ICT on AI safety research and cybersecurity. One concrete piece: the Korea AI Safety Institute will conduct Korean-language model safety evaluations in collaboration with Anthropic. The announcement was updated on June 18 to include those details.
Enterprise adoption is the other headline. NAVER deployed Claude Code to its entire engineering organization, per Anthropic's announcement. Nexon engineering teams use the same tool for live-service game development. LG CNS is rolling Claude out to thousands of employees and intends to extend it across LG Group. Samsung SDS deployed Claude, including Claude Cowork and Claude Code, to employees at Samsung Electronics for knowledge work and agentic workflows. Hanwha Solutions took a different route, bringing Claude to global employees through AWS Bedrock to satisfy in-region data-residency requirements.
Channel Corp rounds out the commercial picture. Its product Channel Talk, a customer AI platform, runs on Claude and is used by over 230,000 companies across Korea, Japan, and the United States.
Research access also forms part of the package. Anthropic will give up to 60 researchers at the National AI Research Lab (NAIRL) access to Claude. NAIRL draws its membership from KAIST, Korea University, Yonsei University, and POSTECH, with work planned on safety, alignment, and model robustness.
Why it matters
Outside the United States, this is the densest cluster of named enterprise deployments Anthropic has announced. Three of the five enterprise partners are IT services arms of major Korean conglomerates, each with large internal headcounts and significant client-facing operations. That combination matters because it signals domestic adoption at scale rather than a handful of pilot contracts.
Korea's position as a Claude market contributed to the move. The country sits among the top dozen globally for Claude.ai usage, per Anthropic's Economic Index. Concentrated developer activity, particularly in technical work, gave Anthropic a genuine reason to plant a permanent office rather than manage the region through a travelling sales team.
The government MOU is also notable. Earlier big enterprise deals like KPMG or PwC carried no formal regulatory dimension. A joint safety-evaluation track with the Korea AI Safety Institute is a different kind of commitment, one that could eventually produce published results on Korean-language model behavior. Whether it does remains to be seen.
All productivity claims in Anthropic's announcement are attributed to partner executives and have not been independently verified.
What to watch next
Seat counts and contract terms were not disclosed for any of these deployments. The practical question is whether organizations like Samsung SDS or LG CNS expand from employee productivity use into customer-facing Claude applications in the next six to twelve months. The NAIRL research track is a longer-horizon signal: any published alignment or safety papers from the 60 affiliated researchers would tell the story of whether Anthropic's academic access program translates into substantive research output.
Sources
- Anthropic: Seoul office and Korean AI ecosystem partnerships (primary, Anthropic announcement June 17-18, 2026)
- Anthropic: KiYoung Choi named Representative Director (secondary, leadership announcement)