[BODY]
U.S. artificial intelligence company Anthropic will run its latest AI model, "Claude Opus 4.8," in a data center located in South Korea. The move lays the groundwork to make a full-scale push into the country's financial and medical sectors—including banks and hospitals—where adoption of foreign AI has been limited due to regulations on sensitive information.
According to the IT industry on the 22nd, Anthropic will within weeks launch an "in-region" service that runs Claude Opus 4.8 in Amazon Web Services' (AWS) Seoul region data center. An Anthropic official said, "We plan to offer Claude Opus 4.8 in-region in the AWS Seoul region within weeks," adding, "We will be able to sell Claude-based AI services even to industries subject to strict data retention regulations."
In-region is a method that confines an AI model's inference to data centers within a single region. Foreign AI services typically exchange data by jointly using data centers around the world. By contrast, in-region keeps the AI model's operation and data processing within the relevant region and restricts sharing with data centers in other regions. If Anthropic stores data and runs AI only in the Seoul region, the export of sensitive information overseas is blocked.
This targets data regulations in finance and healthcare. Under electronic financial supervision rules, financial firms must keep their servers domestically when handling sensitive information such as personal credit information and unique identifying information via cloud systems. Electronic medical record standards also stipulate that when medical institutions use the cloud, the physical location of servers must be limited to within the country. In both sectors, cloud use itself is permitted, but a domestic server is a prerequisite for handling sensitive information.
If Anthropic supports operation in the Seoul region, it opens the way for these institutions to process even sensitive information with Claude. Anthropic plans to meet these requirements and expand its customer base into regulated industries such as finance and healthcare.
As Anthropic's case of breaking through regulations shows, the scope of foreign AI services' penetration into the domestic market is expected to broaden going forward, reaching not only consumers but also corporate and public domains. As AI moves beyond a personal work-assistance tool to aim for a role as an essential asset for governments and companies, the importance of securing AI technology sovereignty is being highlighted accordingly.
Among domestic industries, finance and healthcare are considered the most difficult fields for foreign AI companies to break into, because they handle the most sensitive personal information. Anthropic's planned Seoul in-region support is read as a determination to clear such high barriers and leave no gap anywhere in the Korean business-to-business (B2B) market. Moreover, with OpenAI, which arrived earlier, posting visible results in sales to large corporations and the public sector, a fight for leadership among global AI companies over the Korean market appears inevitable.
The Seoul in-region support was pursued in tandem with the formal opening of Anthropic's Korea office. Anthropic appointed a Korea representative this month, fully kicking off its push into the domestic B2B market. Chris Ciauri, head of Anthropic International, and Choi Ki-young, Anthropic's Korea representative, said at a press conference on the 17th that they are reviewing data retention and processing options tailored to Korea's regulatory environment.
This timing also coincides with the point at which the U.S. administration fully blocked foreign access to Anthropic's next-generation Claude models, such as "Fable5" and "Mythos5." Although the U.S. government took the unprecedented step of imposing export controls on AI models, Anthropic has not bent its stance on global business expansion. Ciauri assessed the Korean market, saying, "Korea is a country with a strong national AI vision and AI regulations, which aligns well with Anthropic's founding goal of providing humanity with a safe AI experience."
Taken together, the Seoul in-region support can be seen as a move to step first into regulated industries and preempt the experience of using Claude. The core targets are finance and healthcare. The two fields must comply with cloud server usage guidelines under electronic financial supervision rules and electronic medical record management standards, respectively, and both systems limit the physical location of cloud servers used by financial firms and medical institutions to within the country.
Google, leading with Gemini, entered the B2B contract battle a step later than Anthropic and OpenAI. Last year it officially launched Gemini Enterprise, an AI platform for businesses. While it has not yet marketed aggressively, it has succeeded in securing client companies. Samsung Electronics adopted Gemini alongside ChatGPT and Claude as generative AI for employees' work, and KakaoBank and CJ Olive Young are also attempting an internal AI transformation (AX) with Gemini Enterprise. Though a latecomer to B2B, Google touts its integration with Google Workspace, a corporate collaboration tool, and Google Cloud Platform (GCP) as strengths.
For this reason, if a financial firm or medical institution entrusts the processing of sensitive information to a global AI service that uses overseas data centers, it constitutes a regulatory violation. This is because global AI companies, including Anthropic, provide services using data centers scattered around the world. An official at a domestic AI model developer said, "In the financial and medical fields, due to IT systems and information storage regulations, there are many cases of giving priority to adopting AI services based on domestic data centers."
OpenAI, which planted its flag in Korea first last year, is expanding its customer base in both the B2B and public domains. In corporate sales, it has signed partnership agreements with major systems integration (SI) companies such as Samsung SDS, LG CNS, and SK AX, selling enterprise ChatGPT through them. LG Electronics, GS Engineering & Construction, Krafton, and Hana Tour, among others, have adopted and are using it.
What is notable is that OpenAI is also putting effort into the public sector. OpenAI recently signed memorandums of understanding (MOUs) with the Korea Water Resources Corporation and the Korea Technology Finance Corporation. The Korea Water Resources Corporation is reviewing ways to preemptively respond to natural disasters such as floods and droughts using OpenAI's technology, while the Korea Technology Finance Corporation is reviewing ways to use AI to evaluate startups' business models.
OpenAI's pioneering of public-sector business is interpreted as a strategy aimed at securing long-term customers rather than immediate revenue. The head of one AI service startup analyzed, "It is aiming for a lock-in effect, instilling the experience of using ChatGPT early on in millions of employees at public institutions and state-owned enterprises nationwide, so that they continue to prefer OpenAI's tools thereafter."
