Anthropic anchors Claude in Korea

Anthropic has announced the opening of its Seoul office and a set of partnerships across South Korea, with a June 18 update adding details on a memorandum of understanding with Korea’s Ministry of Science and ICT. The central fact has two parts: the company is giving Claude a permanent local base in Korea, and it is committing to work with the government on responsible public-sector AI adoption, Korean-language model safety, and AI-enabled cyber threats.

This is not a model launch. It is a deployment signal at a time when leading AI labs are trying to turn assistants into infrastructure for companies and public institutions. South Korea is a logical test case. Anthropic says the country is among the top dozen markets globally for Claude.ai usage, with activity concentrated in technical and creative work. The company also points to an active developer community, Claude meetups running since September 2025, and build events around Claude Code.

The most concrete part is the list of uses Anthropic names. NAVER has deployed Claude Code across its entire engineering organization. Nexon uses Claude Code to write, review, and ship code for live-service games. LG CNS is rolling Claude out to thousands of employees and plans to extend it across LG Group. Hanwha Solutions is using Claude through AWS Bedrock to meet data-residency and security requirements. Samsung SDS is deploying Claude, including Claude Cowork and Claude Code, to teams connected with Samsung Electronics. These names do not prove productivity gains by themselves, but they show adoption moving beyond isolated pilots.

The public-sector agreement adds a different layer. With the Korea AI Safety Institute, Anthropic plans to evaluate model safety in Korean and exchange information on AI-enabled cyber threats. That matters because localizing a model is not just a matter of translating the interface. Risks, data, administrative use cases, and safety tests also have to fit a language and a regulatory context. For governments, the practical challenge is to benefit from existing frontier tools without treating them as neutral imports that can be dropped into public workflows unchanged.

This brief is distinct from recent Anthropic items already covered on Mindshot. It is not about the company’s confidential IPO filing, the Public Record survey, or the suspension of Fable 5 and Mythos 5. The useful information here is operational: Claude is being anchored in an advanced Asian market through a mix of large companies, startups, academic research, and public cooperation on safety. The story to watch next is whether these partnerships produce measurable deployments, and whether the safety work becomes a repeatable pattern for other countries negotiating with AI labs.