Anthropic Opens Seoul Office, US AI Export Controls a Big Shadow

JAKARTA - Anthropic officially opened an office in Seoul, South Korea. The US-based AI company wants to expand its global business, but this move comes as Washington tightens foreign access to advanced AI models.

According to a Yonhap report, Thursday, June 18, Anthropic's expansion into South Korea came after the company revoked access to its two most advanced AI models, Fable 5 and Mythos 5. The decision followed Donald Trump's administration's order to suspend foreign citizens' access to the model.

The Washington Post reported that the policy emerged after Washington officials found that one of the potential recipients of Mythos access was a South Korean telecommunications company that the Trump administration suspected of having ties to China.

"Korea is one of our fastest growing markets in the world because of the technical and developer base you have," said Chris Ciauri, Anthropic's International Managing Director, at a press conference, Wednesday.

He said South Korea has the opportunity to become one of the fastest Anthropic markets in the world.

South Korea ranks 12th out of 116 countries in the use of Claude per capita, according to an Anthropic report published in March. Claude is an AI model made by Anthropic.

"I wouldn't be surprised if the 12th rank soon goes into the single digits," said Ciauri.

As part of its expansion in South Korea, Anthropic plans to build a partner ecosystem with a number of major hyperscalers, including Google Cloud, Microsoft, and Amazon.

Hyperscalers are large-scale cloud computing service providers that are able to handle very large amounts of data and computing needs.

Anthropic is also reviewing local data storage options to comply with rules in South Korea.

Yonhap also reported that Ciauri did not provide further explanation when asked about Washington's restrictions on export controls on the advanced AI model Anthropic.

"We will not comment on Project Glasswing at this time," he said.

Project Glasswing is Anthropic's cybersecurity initiative that gives important organizations and companies early access to the company's advanced AI models.