Claude overtakes ChatGPT in South Korea paid AI market

For the first time, Anthropic's Claude has pulled ahead of OpenAI's ChatGPT when measured by paid spending in South Korea's generative AI sector. The milestone reported by KED Global, points to a growing preference among Korean businesses for enterprise-focused AI tools, even as ChatGPT continues to dominate consumer adoption by a wide margin.

What the Card Payment Data Reveals

Hankyung Aicel, a data analytics division under the Korea Economic Daily group, tracks aggregated credit and corporate card transactions to gauge spending patterns across AI platforms. According to a report by The Korea Herald, Aicel's February figures showed that Korean users collectively spent a record 90.55 billion won (around $60.6 million) on leading generative AI services through card payments.

Claude's share reached 27.12 billion won, crossing the 30 percent threshold for the first time. Corporate cards drove 61 percent of those transactions, a ratio that exceeded ChatGPT's by roughly 16 percentage points. Per-transaction spending on Claude also averaged over 100,000 won, nearly twice the sector-wide norm, while card-linked revenue grew 830.8 percent compared to the prior year.

ChatGPT moved in the opposite direction. Card-based spending on OpenAI's service dropped 7.7 percent from the previous month, extending a slide that began after peaking at 55.8 billion won in November.

Why the Numbers Tell an Incomplete Story

OpenAI Korea's general manager, Kim Kyoung-hoon, challenged the methodology publicly on LinkedIn. His argument: card data structurally undercounts enterprise contracts because many large organizations pay through invoice billing, not corporate cards. When companies upgrade from ChatGPT Business to the higher-tier Enterprise plan, they typically switch payment channels entirely, making it look like a spending decline when it may simply reflect a billing migration.

There is evidence supporting that view. Samsung SDS, which secured reseller rights for ChatGPT Enterprise in South Korea last December, recently confirmed partnerships with over ten corporate clients spanning manufacturing, finance, and services. None of those agreements would show up in card transaction records.

On raw user volume, ChatGPT still leads overwhelmingly. Data from WiseApp showed 22.93 million monthly active users for ChatGPT in February across South Korea, while Claude registered just 770,000. The gap highlights how these two platforms serve fundamentally different audiences: ChatGPT attracts individual users and small teams, while Claude is carving out territory among higher-spending corporate buyers.

South Korea as a Proving Ground for Enterprise AI

Few markets carry as much weight in the paid AI market as South Korea. According to Sensor Tower, the country generates more cumulative ChatGPT app revenue than any nation except the United States, despite having a relatively small install base. Anthropic's Economic Index from January 2026 ranked South Korea seventh globally for Claude usage relative to population size, placing it within the top quartile worldwide.

Both companies are now expanding their Korean operations aggressively. OpenAI launched its Seoul office last September and has been building out a dedicated sales operation. Anthropic, meanwhile, is staffing up in Seoul as it prepares to open its third office in the Asia-Pacific region, after Tokyo and Bengaluru.

Globally, the enterprise battle between these two firms is tightening. Menlo Ventures estimated last December that Anthropic controlled roughly 40 percent of the enterprise LLM API market, with OpenAI at 27 percent. South Korea's spending data reinforces that trend: even where ChatGPT holds massive consumer reach, Claude is capturing the higher-value contracts that increasingly define where AI revenue actually comes from.