Anthropic Opens Seoul Office, Lands NAVER, Samsung, LG, and Nexon as Claude Deploy Partners

There are few contexts stranger for a corporate office opening than having your flagship models banned from foreign national access six days earlier. Yet that is precisely the circumstance Anthropic found itself in when it formally launched its Seoul office on June 17 — and the company’s decision to proceed, and the scale of the Korean enterprise partnerships it unveiled simultaneously, tells you something about how seriously the AI market has taken root in South Korea.

Anthropic’s Seoul office is its third in Asia-Pacific, after Tokyo and Bengaluru, and its ninth globally. KiYoung Choi, former General Manager of Snowflake Korea, leads the office as Representative Director. Alongside the opening, Anthropic announced enterprise deployments, research partnerships, government agreements, and developer programs covering virtually every tier of South Korea’s technology ecosystem.

The Enterprise Wave

The scope of simultaneous deployments is striking. NAVER, South Korea’s dominant internet company and the closest Korean equivalent to Google, deployed Claude Code across its entire engineering organization — a decision that would have taken months of internal evaluation to reach. Samsung SDS, the IT services arm of Samsung Group, is rolling out Claude Code and Claude Cowork across Samsung Electronics, targeting both software development workflows and knowledge work. LG CNS, the IT subsidiary of LG Group, is deploying Claude to thousands of employees for software development and client solutions.

Nexon, one of South Korea’s largest gaming companies and the creator of MapleStory and FIFA Online, is using Claude Code across its engineering teams for live-service game development. Hanwha Solutions is deploying Claude via AWS Bedrock with strict in-region data controls — a configuration designed for companies with regulatory constraints around data residency. Channel Corp, operator of Channel Talk, is using Claude to power its customer communication platform serving more than 230,000 companies across Korea, Japan, and the United States.

The breadth of adoption across sectors — search, electronics, gaming, energy, and B2B SaaS — suggests that Korean enterprises have been evaluating Claude for longer than this week’s announcements imply. Enterprise software deployments at this scale don’t happen in six days.

Government and Research Partnerships

Alongside enterprise deals, Anthropic signed a Memorandum of Understanding with South Korea’s Ministry of Science and ICT covering Korean-language AI safety evaluations and cyber threat information sharing. The agreement provides Claude access to sixty researchers affiliated with the National AI Research Lab consortium — KAIST, Korea University, Yonsei University, and POSTECH — who will work on AI safety, model evaluation, alignment, and robustness.

The government MOU is significant given the backdrop. South Korean participants in Project Glasswing, including Samsung Electronics and SK Hynix, had their access revoked following the June 12 export control directive. The new bilateral agreement appears designed to create a formal governmental channel for model access and safety research that doesn’t depend on the informal program structures that triggered scrutiny.

Anthropic also announced nonprofit deployments: Good Neighbors Korea, a child rights organization, is using Claude to analyze program outcomes and reduce administrative burden for its staff.

The Surreal Backdrop

The office opening was overshadowed, and in some ways defined, by the ongoing Fable 5 and Mythos 5 access ban. Chris Ciauri, Anthropic’s Managing Director of International, addressed the situation directly at the opening event, expressing confidence that “in the coming days, the models will become available again.” He framed the Seoul launch as a demonstration of Anthropic’s long-term commitment to the Korean market regardless of the short-term disruption.

The companies that deployed Claude for Korean enterprise use are currently restricted to the Claude 3 series models — the older generation predating Fable 5 and Mythos 5. If and when the export restrictions are lifted, those deployments are positioned to immediately upgrade.

The timing also carries IPO implications. Anthropic is reportedly preparing for a public offering that would value the company in the range of $852 billion to nearly $1 trillion. South Korea’s technology sector, which includes some of the world’s largest chip manufacturers and internet companies, represents both a significant revenue opportunity and a credibility signal for institutional investors evaluating whether Anthropic’s commercial traction justifies its valuation.

Developer Ecosystem

Beyond enterprise and government, Anthropic launched the Claude for Startups program in Korea, partnering with BASS Ventures to reach early-stage Korean AI companies. The company co-hosted Claude Build Day and a Push to Prod hackathon with Replit and Korean investment partners, signaling an intent to cultivate the developer ecosystem rather than rely solely on top-down enterprise sales.

The developer-focused events attracted significant attention in Seoul’s growing AI startup scene, where Claude has become a preferred model for inference-heavy applications due to its long context window and instruction-following quality.

What It Means for the Region

Anthropic’s simultaneous entry into South Korea with six major enterprise deployments, a government MOU, and a university research partnership is not a standard office-opening playbook. It reflects a company that has spent considerable time cultivating Korean relationships behind the scenes and chose the Seoul office launch as the moment to surface all of them at once.

South Korea is one of the few countries with both the technological sophistication to deploy foundation models at scale and the manufacturing weight — in semiconductors, displays, and hardware — to matter strategically to AI infrastructure providers. The country’s major firms are simultaneously potential customers, research partners, and components of the global chip supply chain that Anthropic depends on.

For Anthropic’s competitors, the Korean market entry is a competitive signal that warrants a response. Google DeepMind, OpenAI, and Microsoft have all been active in Korea, but Anthropic’s breadth of simultaneous partnership announcements suggests it has been executing a coordinated market-entry strategy rather than reacting opportunistically.

The office opens amid a ban on Anthropic’s most advanced models. That it opened anyway, with this many partners in tow, is the clearest possible signal of where both Anthropic and South Korea’s technology industry see the long-term relationship heading.