Claude's Surge in South Korea's Enterprise AI Market Ignites a Public Dispute With OpenAI

A detailed card-spending analysis showing Claude rapidly gaining ground on ChatGPT in South Korea’s corporate market prompted a rare public rebuttal from OpenAI’s Korea chief — revealing how high-stakes the enterprise AI race has become in one of the world’s most commercially active AI markets.

Key Takeaways

• Claude hit 30% of card-based AI spending in Korea: According to card transaction data from Hankyung Aicel, Claude accounted for 27.12 billion won (~$18.2M) out of 90.55 billion won in total card spending on generative AI services in South Korea in February 2026 — a first-time threshold crossing. Claude’s card spending surged 830.8% year-over-year.
• OpenAI Korea disputes the methodology: General Manager Kim Kyoung-hoon publicly argued that card data structurally undercounts large enterprise contracts, which are typically settled via invoice. He noted that ChatGPT Enterprise clients specifically shift from card to invoice billing when upgrading, making declines in card spend look worse than they are.
• The data tells two different stories: Claude’s average transaction exceeded 100,000 won — roughly double the industry average — and corporate cards made up 61% of its payments, 16 percentage points above ChatGPT. But ChatGPT still leads in users: 22.93 million monthly active users vs. 770,000 for Claude. The split reflects a broader global pattern where ChatGPT dominates consumer volume and Claude concentrates on higher-value enterprise use.

South Korea ranks seventh globally in Claude usage intensity according to Anthropic’s January 2026 Economic Index, and second globally in cumulative ChatGPT app revenue. Anthropic is actively expanding its Seoul office presence. Menlo Ventures estimated in December that Anthropic holds 40% of the global enterprise LLM API market versus 27% for OpenAI.

Read the full article on The Korea Herald