The government has moved to expand its artificial intelligence (AI) cooperation structure, which had centered on OpenAI, by pursuing policy cooperation with Anthropic. The move is seen as an attempt to diversify cooperation partners as the global AI competitive landscape shifts.
On the 15th, according to the government and industry, the Ministry of Science and ICT is reviewing official cooperation measures with Anthropic, which developed the Generative AI model "Claude," including signing a memorandum of understanding (MOU).
The discussions are said to have begun after Bae Kyung-hoon, Deputy Prime Minister and Minister, met with Anthropic Chief Executive Officer (CEO) Dario Amodei at the "2026 Artificial Intelligence (AI) Impact Summit" held in New Delhi, India, on the previous month.
At the time, Bae discussed Anthropic's plan to open a Seoul office and the potential for cooperation, and reportedly exchanged views on trends in the software-as-a-service (SaaS) market that have changed since the emergence of Claude Code.
Anthropic had earlier announced plans to establish a Korea office in Gangnam, Seoul, early this year, but with the selection of a country head still underway, the signing of an official cooperation agreement with the government is said to be somewhat delayed.
The government is particularly focusing on the fact that Anthropic has emphasized the safe development and use of Generative AI. Using the cooperation framework between the U.K. AI Safety Institute and the Korea AISI (Korea Artificial Intelligence Safety Institute), it is also reviewing ways to expand joint research related to AI safety.
Until now, the Ministry of Science and ICT's global AI cooperation had been promoted with a focus on OpenAI. When OpenAI CEO Sam Altman visited Korea last year, the government signed a memorandum of understanding and agreed to cooperate on building the domestic AI ecosystem, AI transition in the public institutional sector, and AI talent development.
In addition, it has pushed cooperation on AI infrastructure expansion by agreeing to participate in "Stargate," a large-scale data center construction project that OpenAI is pursuing with SoftBank and Oracle.
However, as Anthropic has grown rapidly in recent months, the government's global AI cooperation strategy appears to have shifted. According to research institute Epoch AI, after surpassing $1 billion in annual revenue, Anthropic has grown about tenfold each year, far outpacing OpenAI's annual growth rate.
Unlike OpenAI, which has taken a strategy centered on a chatbot for general users, Anthropic is also assessed to have built a stable revenue base by focusing on advancing business-to-business (B2B) AI models.
Within the government, concerns have been raised that concentrating cooperation on specific corporations could be disadvantageous in processes such as public procurement, AI standards, and regulatory negotiations.
However, in sensitive areas such as defense and healthcare, the approach of prioritizing the "independent AI foundation model project" that the government is supporting is expected to remain in place. The government is pursuing a two-track strategy that combines a "sovereign AI" strategy to enhance Korea's AI technological competitiveness with cooperation with global corporations.
Meanwhile, Korea is also considered an important market for Anthropic. According to company data, Korea ranks in the global top five in both total Claude usage and per capita usage, and domestic developers are among the groups that use the AI coding tool "Claude Code" the most worldwide.