三星电机CAIO受访警示中国AI零部件追赶并推进内部AI转型，属战略叙事无新增可交易事实

Samsung Electro-Mechanics (009150) is expected to post unprecedented earnings growth this year as orders surge for components used in artificial intelligence chips supplied to Big Tech. Inside the company, however, concern is mounting that it could lose its market lead at any time to Chinese rivals, who are closing the gap quickly in AI components on the strength of sheer numbers. In response, the company has created a chief AI officer post and held an in-house AI hackathon as part of a push to widen its technology gap through AI transformation.
Jang Woo-suk, executive vice president and chief AI officer at Samsung Electro-Mechanics, met with The Seoul Economic Daily at the company's Suwon plant on the 13th and said the company cannot avoid what he called a "red sandwich crisis," in which China catches up rapidly even as Samsung Electro-Mechanics expands quickly on the AI boom. "There are more than 10 times as many companies in the same business in China," the CAIO said. "That means we cannot hold off the pursuit by developing technology at the same pace they do."
Samsung Electro-Mechanics is projected to report revenue of 14.43 trillion won ($10.1 billion) and operating profit of 2.05 trillion won ($1.4 billion) this year, according to financial data provider FnGuide. That would be up 28% and 124%, respectively, from a year earlier. The gains are driven by sharply higher sales of multilayer ceramic capacitors and flip-chip ball grid array substrates used in Big Tech's AI chips and related infrastructure.
Even with that performance, the company's first CAIO argues that only pre-emptive preparation for China's advance can prevent an erosion of its market position later.
The company created the CAIO post in August and appointed Jang to the role to drive its internal AI transformation. The red sandwich crisis he described refers to a two-front squeeze from Chinese companies, on advanced technology on one side and price competitiveness on the other.
"We have assigned AI transformation projects to every team as a requirement, and more than 100 projects are under way," the CAIO said. "We plan to expand projects related to physical AI that can be applied on the manufacturing floor in particular." Another core task is applying the know-how accumulated by veteran engineers to physical AI and robots, while building "AI-ready data" so that complex data from the factory floor can be used efficiently.
Samsung Electro-Mechanics developed its own AI model, SEMCo-Work, and held its first hackathon at headquarters on the same day to spread its use across the company. SEMCo-Work was built on the company's own servers and security systems. Over four days through the 16th, 200 employees will use SEMCo-Work to take on the task of developing agentic AI that cuts repetitive work and can be applied directly to their jobs.
Participants will build and submit tools they can use in their own work, reflecting the nature of a manufacturer whose operations range from product development to production, quality control and customer response. Prize money for the gold, silver, bronze and honorable mention teams totals 33 million won. The AI agents selected will be shared internally so that all employees can use them.
"The AI era is still only Jan. 1, and we must respond actively to the growth and innovation that AI is driving," said Jang Duck-hyun, president of Samsung Electro-Mechanics. "The experience of building an agent suited to your own work through this hackathon will be the start of 'one person, one AI agent.'"