Skip to main content

Samsung seeks to expand from AI chips into data-center infrastructure

15 NOVEMBER 2026·2 MIN READ·6 SOURCES

Samsung is seeking to expand beyond semiconductor manufacturing into AI infrastructure, including data-center construction and operations, according to a report by Finansavisen citing the Financial Times. Its reported moves include a $1 billion investment in Helix Digital Infrastructure and a purchase of a European cooling-systems manufacturer.

Key takeaways · 4

  • 01

    Samsung is aiming to extend its AI business from memory chips into data-center infrastructure and operations.

  • 02

    Samsung SDS’s Gumi facility is planned to start at 60 megawatts and expand in phases to a maximum of 120 megawatts.

  • 03

    Samsung SDS is in preliminary discussions with OpenAI and Anthropic about becoming data-center customers; the discussions are not reported as completed deals.

  • 04

    Analysts estimate Samsung SDS could generate 6.6 trillion won ($4.9 billion) in annual data-center revenue by 2031.

A broader infrastructure strategy

Samsung has benefited from demand for memory chips used in artificial intelligence, and now wants to expand beyond semiconductor manufacturing into AI infrastructure, Finansavisen reported, citing the Financial Times.[1] The company’s stated goal is to provide products and services spanning memory chips, batteries, and data-center construction and operations.[1] Finansavisen also reported that Samsung invested $1 billion in Helix Digital Infrastructure, bought a European cooling-systems manufacturer, and entered into a partnership with OpenAI.[1] Samsung Electronics supplies memory chips and cooling systems, while Samsung C&T could build data centers.[1]

Gumi facility adds a concrete project

Samsung SDS announced on November 2 that it had broken ground on a dedicated AI data center in Gumi, South Korea.[2] The site in Gongdan-dong covers approximately 46,000 square meters, and the facility is planned to begin with 60 megawatts of power capacity, expanding in phases to a maximum of 120 megawatts.[2] Samsung SDS scheduled completion for the end of 2028 and disclosed in a January regulatory filing that it would invest 427.3 billion won in the project.[2] The plan includes hybrid air-and-liquid cooling and colocation, allowing customers to install and use their own GPUs.[2]

Customer relationships remain at different stages

Samsung SDS is in preliminary discussions with OpenAI and Anthropic about bringing them on as data-center customers, according to Finansavisen.[1] Separately, Samsung SDS said it signed an OpenAI reseller partnership on December 23, 2025, to offer ChatGPT Enterprise and technical support to Korean enterprise customers.[5] Samsung SDS also said it would become the first South Korean pilot partner in OpenAI’s Daybreak cybersecurity program and established a strategic partnership with Anthropic in San Francisco in July.[6] These announcements describe distinct relationships; the reported data-center customer discussions are preliminary.[1]

Revenue potential and market forecasts

Analysts estimate Samsung SDS could generate 6.6 trillion won, or $4.9 billion, in annual data-center revenue by 2031—about six times its current level, according to Finansavisen.[1] McKinsey forecasts global demand for data-center capacity could nearly triple by 2030, with AI accounting for about 70% of demand.[1] The Gumi facility is planned to support large-scale computing, including AI training for precision-task robots at manufacturing sites.[2] Separately, the Financial Times reported that meeting AI computing demand could require $5.2 trillion in investment.[1] These are forecasts and projections, not reported results.[1]

For professionals evaluating AI infrastructure, Samsung’s plans point to a supplier seeking to connect chips, cooling, construction and operations. Treat projected revenue and prospective customer relationships as planning signals rather than secured outcomes, and distinguish Samsung SDS’s announced services from the broader group strategy.

Why it matters
For teams

AI training for your team

kju.ai is built for teams: daily sessions, team analytics, and rollout support for your organization.

How this developed

  1. 15 November 2026

    Samsung seeks to expand from AI chips into data-center infrastructure

Sources

AI fluency, one session a day, built for your work.