Anthropic Opens Seoul Office, Names KiYoung Choi to Run Korea

Anthropic announced on May 26 that it has appointed KiYoung Choi as Representative Director of Anthropic Korea, the legal title that effectively makes him the country CEO. The Seoul office opens shortly after, with senior leadership from Anthropic HQ traveling to Korea in the next few weeks to formally launch the operation and meet major customers.

Why this matters

Korean users are already running Claude at 3.5 times the rate that a country of 52 million people would predict, according to Anthropic's March Economic Index. That number is one of the largest population-adjusted usage gaps Anthropic has disclosed for any market. The Seoul office isn't an experiment. It's catching up to revenue that's already on the books.

A serious hire, not a placeholder

Choi brings 30 years of enterprise software experience in Korea — a country where local relationships and government access decide which platforms get adopted in regulated industries. Most recently he served as Country Manager of Snowflake Korea. Before that, he held leadership roles at Google Cloud Korea, Adobe Korea, Autodesk Korea, and Microsoft Korea, where he served as Chief Operating Officer.

The pattern: every previous employer was a US enterprise software company trying to crack Korea. Anthropic skipped the usual rookie phase and hired someone who has done this exact job four times.

The customer base is already there

Korean conglomerates are heavy Claude adopters. Samsung, SK Telecom, LG, Hyundai, and Kakao all run AI workloads at scale. SK Telecom has integrated Claude into customer-facing AI products. Law&Company, the largest legal tech firm in Korea, uses Claude for legal workflow automation. The Korean usage skew isn't a curiosity — it's concentrated in high-difficulty technical R&D and creative work, the categories where Claude Opus 4.7 and 4.8 perform best.

According to Anthropic, the Korea team will focus on three lines: enterprise and startup partnerships, government and research-institution engagement, and developer community support. That third one matters more than it sounds. Korean developers are over-indexed on Claude Code adoption, and a local office means Korean-language documentation, support hours that match the time zone, and on-the-ground evangelism at conferences like SK AI Summit and Samsung Developer Conference.

The Mythos angle

There is a security dimension here that the press release does not state plainly. The Elec, a Korean semiconductor and electronics publication, reported that Anthropic had already conducted cybersecurity cooperation discussions using Claude Mythos Preview at a workshop with Korean government agencies in May. The agencies involved: Ministry of Science and ICT, Ministry of Foreign Affairs, National Intelligence Service, Financial Services Commission, National AI Strategy Committee, and Financial Security Institute.

That is not a normal sales meeting. That is Anthropic briefing South Korea's intelligence and financial-regulator establishment on Mythos's capabilities — and those agencies apparently deciding the technology is worth a cooperative relationship. Industry sources cited by The Elec expect Korean conglomerates and cybersecurity firms to gain controlled access to Mythos and the Claude Security public beta once the Seoul office formally launches under Choi.

Geopolitical context

South Korea sits on the front line of North Korean state-sponsored cyber operations. The National Intelligence Service has documented years of attacks against Korean financial, nuclear, and defense systems. A tool that can autonomously surface critical vulnerabilities in legacy infrastructure — which Project Glasswing has demonstrated Mythos can do — is not a productivity tool in this context. It is a national-security capability.

That framing helps explain why Korean government engagement preceded the office opening rather than following it.

What it means for global expansion

The Seoul office is Anthropic's sixth international location, joining London, Tokyo, Bengaluru, Singapore, and Milan (opened the day after Seoul, on May 27). Two offices opening within 48 hours signals that the company is no longer rationing international expansion. It's racing OpenAI's enterprise sales team to lock in the next wave of major accounts.

Compared to OpenAI's international presence, Anthropic is now broadly comparable in office count but more aggressive in 2026 — six international hubs added in roughly 18 months, with the Korea and Italy openings on consecutive days.

What to watch

The first concrete signals from the Seoul office will be (1) the named customer logos that Anthropic discloses in the next earnings disclosure or analyst call, (2) whether Mythos Preview access lands at a Korean chaebol cyber team before any other Asian customer, and (3) Korean-language pricing and SLA documentation that would indicate a self-serve product launch.

For now, the office is a recognition of where Claude already runs. Korea was Anthropic's biggest unserviced market. That ends this week.

Bottom line

A country-CEO hire from Snowflake, a one-day-later Milan opening, and a sitting workshop with Korea's national-security apparatus are three different signals pointing the same direction. Anthropic is operationalizing markets it had been treating as remote billing addresses. Korean enterprises that wanted local Anthropic support without going through Tokyo or Singapore now have it. The Mythos pipeline that runs through this office is the more strategic story, and it is the one to track for the back half of 2026.