AI apps including Claude are displayed on a screen. [Getty Images Bank]
AI apps including Claude are displayed on a screen. [Getty Images Bank]

The Claude app recorded about 2.41 million monthly active users in South Korea in April, according to market research firm WiseApp Retail.

Anthropic's AI model Claude hit a record high market share in South Korea's chatbot market last month, as the company's push into the domestic market — including the official launch of a local subsidiary — appears to be paying off. Industry observers say Anthropic has grown rapidly by targeting not general consumers but the corporate AI transformation, or AX, market.

Claude's share of South Korea's chatbot market reached 5.75% last month, an all-time high, according to market research firm StatCounter.

Industry officials attribute the gain to Anthropic's aggressive domestic push, which included opening a Korea office last month. Claude's market share stood at just 0.9% as recently as January before climbing to 4.77% in April. That same month, Claude's monthly active users surged to about 2.41 million — up sharply from 530,000 the previous month, according to WiseApp Retail. The model then hit its record high in June, when Anthropic formally launched its Korea subsidiary.

A separate dataset points to a similar trend. AI startup MindLogic analyzed domestic generative AI usage patterns based on data from its AI agent platform FactChat, finding that Claude surpassed GPT in token usage share for the first time in May. Within FactChat, GPT models' share of token usage fell from 85.7% in September last year to 34.8% in May, while Claude's share surged from 5.7% to 36% over the same period.

Choi Ki-young, head of Anthropic Korea, speaks at a press conference marking the launch of the company's Seoul office. [Anthropic]
Choi Ki-young, head of Anthropic Korea, speaks at a press conference marking the launch of the company's Seoul office. [Anthropic]

Industry insiders credit Anthropic's rapid growth in South Korea to its strategy of targeting the corporate AX market. When Anthropic officially opened its Korea office last month, the company confirmed it would pursue the enterprise AI market with Claude and its AI coding agent Claude Code.

Choi Ki-young was named the inaugural head of Anthropic Korea. Choi is a B2B and cloud sales specialist who previously served as head of Google Cloud Korea and as a vice president at Microsoft Korea. The industry views the appointment as consistent with Anthropic's strategy of targeting corporate clients rather than general consumers.

At a press conference last month marking the Korea office opening, Choi said South Koreans use Claude for work more than people in other countries. "We will think about how to tailor our services for Korea and support the Korean business and developer community," he said.

The rapid rise of second-tier AI models — including Claude and Google's Gemini — is shaking ChatGPT's dominance in South Korea. ChatGPT's domestic chatbot market share exceeded 80% as recently as October last year but has been declining steadily this year, according to StatCounter. It fell to 56.58% in April before recovering slightly to 66.58% in June.

With its position under pressure, OpenAI has also moved to reclaim domestic market share by targeting the large-enterprise AX market. The industry now expects OpenAI and Anthropic to clash head-on in that space.

Anthropic's headquarters in the United States. [AP]
Anthropic's headquarters in the United States. [AP]

Earlier in June, OpenAI struck a large-scale enterprise deal with Samsung Electronics to provide ChatGPT Enterprise and Codex to Samsung employees. Anthropic, meanwhile, plans to expand its footprint by supplying Claude to leading Korean IT companies including Naver and Nexon, as well as partnering with large conglomerate-affiliated systems integration firms such as LG CNS.

A potential complication for Anthropic, however, is pressure from the US government. Last month, Washington announced restrictions limiting foreign nationals' access to Anthropic's latest AI models, Mythos 5 and Fable 5 — marking the first time the US government had imposed export controls on a domestic company's own products. The measures were lifted June 30, about two weeks after they were introduced.

The fact that political scrutiny has fallen on Anthropic rather than OpenAI could heighten concerns among South Korean clients about relying on Claude. "For enterprise customers, the stability of supply and regulatory risk matter just as much as an AI model's performance," one industry official said. "Anthropic's technological competitiveness is rising fast, but if US export control issues recur, domestic companies may feel uneasy about applying it to core business operations."


chami@heraldcorp.com