"Speed vs. Infrastructure: Coupang and Shinsegae Wage AI War for Retail Dominance"

[Korea Financial Times, Park seulgi] The retail industry's competitive front is shifting decisively toward artificial intelligence. Domestic retailers are joining forces with global tech giants to build AI infrastructure, fundamentally reshaping the terms of competition.

Coupang has partnered with Nvidia to build an "AI Factory" aimed at upgrading its entire logistics and recommendation systems, while Shinsegae Group is working with Reflection AI to establish what would be the largest AI data center in South Korea. Though their approaches differ, both strategies center on restructuring retail competitiveness around data and AI.

According to industry sources on the 23th, Coupang and Shinsegae Group recently teamed up with Nvidia and Reflection AI, respectively, to enhance their competitiveness through AI. The accelerating investment reflects rapid shifts in the consumer environment, including personalized recommendations, fast delivery, and price competition.
Industry observers say AI has become an essential competitive asset as the factors driving retail competition increasingly shift to a data-driven structure. AI has evolved beyond a simple efficiency tool into a core determinant of corporate competitiveness. The broader trend of collaboration with global tech giants is similarly seen as an effort to secure a stronger edge in the AI race through technology acquisition.

"Retail competition is no longer decided by stores or products — it's decided by data and AI," said one industry official. "How quickly a company builds and deploys its infrastructure will determine how the market shifts."

Coupang Accelerates Logistics and Delivery Innovation

Coupang's parent company, Coupang, Inc., recently announced the establishment of an "AI Factory" developed in collaboration with Nvidia, designed to accelerate innovation across its e-commerce logistics and delivery operations. The strategy calls for applying AI throughout the entire process — from product demand forecasting and inventory placement to delivery route optimization — in order to maximize operational efficiency.

Central to the partnership is the creation of an environment that enables rapid AI model development and experimentation. Coupang integrated Nvidia's high-performance AI servers with its proprietary cloud infrastructure system (CIC), allowing developers across its global operations in South Korea and in Seattle and Mountain View in the United States to simultaneously develop AI models within a unified environment. The company says this has significantly accelerated development speed by enabling AI experimentation and deployment without the need to build separate infrastructure.

Results are already emerging. According to Coupang, the application of AI models has improved warehouse scheduling and product loading efficiency, while GPU utilization for AI computing has risen from 65% to 95%.
The industry expects AI-driven innovation in Rocket Delivery — Coupang's signature same-day delivery service — to substantially improve both speed and efficiency.

Since its founding in 2010, Coupang has invested billions of dollars in AI, robotics, and logistics automation. The company currently operates all major business functions — including inventory management, delivery, and customer recommendations — through AI-based technologies, and plans to further advance these capabilities through its Nvidia partnership.

"Logistics and data are the core of retail, and the gap between companies will widen significantly depending on how quickly AI is deployed on the ground," said one industry official.

Shinsegae Secures AI Competitiveness Through Data Center Infrastructure

While Shinsegae Group shares the same AI-driven orientation, its approach diverges markedly. Where Coupang focuses on rapidly applying AI to logistics and delivery to maximize operational efficiency, Shinsegae has opted to preemptively secure AI infrastructure itself through the construction of a large-scale data center. The conglomerate has partnered with U.S.-based AI firm Reflection AI to build what would be the largest AI data center in South Korea — a strategy aimed at directly controlling AI infrastructure built on retail-derived data, positioning it as an engine of future growth.

Shinsegae and Reflection AI plan to construct a domestic AI data center with a power capacity of 250 megawatts — a scale that would surpass any AI data center currently built or under development in South Korea. The two companies have also secured the foundation for large-scale AI infrastructure by agreeing to source key equipment, including GPUs, from Nvidia.

The project goes beyond mere infrastructure development, targeting what the companies describe as "sovereign AI" — the ability for enterprises and institutions to directly control their data without relying on external parties. The "open-weight AI model" being developed by Reflection AI is designed around this principle and aligns with the South Korean government's policy agenda to strengthen national AI competitiveness. The two companies plan to pursue a "full-stack AI business" that extends from data center operations to cloud services and customized AI solutions.

Through this investment, Shinsegae aims to fully leverage the synergy between its retail operations and AI. The plan is to combine years of accumulated customer data with AI technology to enhance overall retail competitiveness — including personalized product recommendations, demand forecasting, and logistics efficiency. The company also intends to develop an "AI commerce" ecosystem that seamlessly connects AI-powered product recommendations with payment and delivery.

As a data- and AI-driven competitive landscape takes shape across the retail industry, observers expect corporate AI rivalry to intensify further. In particular, collaboration with global tech giants has effectively become a prerequisite, making the race to secure technological capabilities all the more fierce.

"AI is no longer a matter of choice — it's a matter of survival," said one industry official. "Going forward, the gap between retailers will not come down to technology itself, but to how quickly they can execute."

Park seulgi (seulgi@fntimes.com)