South Korea

정지훈 파트너 "CES 2026, AI 유틸리티 시대 서막"

ZD Net Korea | Local Language | News | Jan. 23, 2026 | UndeterminedTech Development/Adoption

At CES 2026, Jihun Jung, founding partner of Asia2G Capital and adjunct professor at DGIST, declared the start of the "AI utility era," marking a shift from AI's flashy performance to practical AI applications that create value. He emphasized that this year’s CES was more focused on B2B rather than consumer markets, noting key exhibition areas such as Hyundai Motor’s presence in the West Hall and humanoid robots in the North Hall, while the traditionally prominent Central Hall saw less attention. Samsung notably opted out of the main exhibition hall and held separate events instead.

Jung identified four major battlegrounds for AI development after 2025: a new computing stack, human-AI interfaces, AI’s expansion into the physical world, and enterprise adoption challenges. He highlighted the growing importance of edge semiconductors amid the transition from inference centered on companies like NVIDIA to broader AI infrastructure. He noted that NVIDIA’s large energy consumption remains problematic compared to Qualcomm and startups like HyperExcel, in which Asia2G has invested.

Regarding future computing devices, Jung pointed to glasses-type devices becoming mainstream around 2029–2030, requiring new operating system and software stacks, with Qualcomm likely providing the standard chips. He critiqued China’s weak OS and software capabilities and predicted Microsoft is at risk due to increasing cloud-based tool adoption, with Google’s ecosystem gaining more prominence among younger users.

Jung observed that AI is extending into physical realms, with autonomous vehicle technology showcased prominently, exemplified by companies such as John Deere in agriculture, Caterpillar in construction, and Brunswick in marine yachts. He noted significant growth in AI and robotics innovation award winners at CES 2025 and 2026. Tesla was described as the "Apple of the physical AI era," with its end-to-end control both a strength and a weakness, while NVIDIA compensates for lacking hardware through partnerships, notably with Hyundai Motor.

The presentation also underscored a strong Hyundai-NVIDIA collaboration, likening it to the Google-Samsung smartphone alliance. Hyundai’s recruitment of former NVIDIA executive Park Min-woo as president overseeing autonomous driving signals deepening ties. Jung downplayed concerns over China’s robotics advances, noting their robots’ limited load capacity compared to Hyundai and Tesla.

Finally, Jung forecasted that AI, semiconductors, robotics, energy, and wearables will converge and drive the metaverse revolution over the next two decades, emphasizing that this multi-technology integration, not any single innovation, will spark transformative change occurring once every 10–20 years.

"물류는 매일 바뀐다…유연한 로봇 자동화가 답"

ZD Net Korea | Local Language | News | Jan. 23, 2026 | Supply Chain Issues

CJ Logistics is extensively implementing a "physical AI" strategy that integrates artificial intelligence, robotics, and simulation to address the dynamic and complex nature of logistics environments. Goo Seong-yong, head of automation development, highlighted that unlike manufacturing, logistics requires flexible use of space and the handling of varying product types throughout the day, which renders traditional fixed automation equipment like conveyors inadequate. The company is therefore focusing on robot-based automation that offers mobility and adaptability.

At e-commerce fulfillment centers such as the Incheon GDC, CJ Logistics has introduced mobile palletizers capable of handling multiple types of boxes in random order, enabled by reinforcement learning-based simulations. Similarly, a random piece-picking robot was deployed at the "The Market" center in Dongtan to manage about 300 product types using improved AI segmentation models, significantly reducing development time through virtual simulation environments.

CJ Logistics prioritizes development speed as a competitive edge, aiming to complete the transition from line construction to operation within three months post-contract. To achieve this, all robot and AI development is based on the Robot Operating System (ROS), with a significant proportion of internal developers trained accordingly. The company leverages virtual data generation and simulation to overcome challenges related to frequent changes in box designs, enabling rapid adaptation within hours.

Further innovation includes a vision-language-action (VLA) model using a dual-arm robot, developed in collaboration with ROBOTIS, which demonstrated quick deployment capabilities. Goo emphasized that despite perceptions of high AI costs, well-trained models can substantially reduce on-site development costs and time, proving cost-effectiveness in logistics automation. He also noted that ROS has unified robotics technology for 18 years and is expected to accelerate change in the logistics sector over the next 5 to 10 years.

Joongang Ilbo | English | News | Jan. 23, 2026 | Regulatory Enforcement Actions

South Korea's Fair Trade Commission (FTC) has fined four major commercial banks—Shinhan Bank, Woori Bank, Hana Bank, and KB Kookmin Bank—a total of 272 billion won ($183.7 million) for allegedly colluding on loan-to-value (LTV) ratios in real estate lending. The banks exchanged internal documents related to LTV ratios and coordinated lending limits from March 2022 to March 2024, restricting competition in the mortgage loan market.

The FTC estimated that the collusion helped the banks generate approximately 6.8 trillion won in interest earnings by reducing uncertainty about competitors’ strategies and avoiding competition on LTV ratios. These four banks collectively hold about 60 percent of South Korea’s real estate mortgage loan market, limiting consumer choice and affecting lending conditions.

The collusion particularly harmed small and midsize enterprises and small business owners, who often have lower credit ratings and rely heavily on secured loans. The case is notable as the first enforcement under a revised fair trade law effective December 30, 2021, which prohibits anticompetitive collusion involving the exchange of sensitive business information.