"Semiconductors, Data Centers, and Physical AI Must Be Integrated"

"Connecting All Three Industries Is Essential for a Nationwide Flywheel"

"Physical AI as the 'Second Semiconductor'... Korea Must Become a Key Player Before Supply Chains Solidify"

Kim Yongbum, policy chief at the Presidential Office, proposed "Project Trinity" on June 11—an initiative that integrates semiconductors, AI data centers, and physical AI as a unified industrial strategy. He explained that in the era of artificial intelligence (AI), competition is shifting beyond individual companies or standalone technologies, and is now evolving into a nationwide effort that combines computing infrastructure, advanced semiconductor supply chains, and the ability to implement AI in manufacturing environments.

Kim Yongbum, policy chief, is attending the K-Chosun Future Vision Meeting presided over by President Lee Jae-myung at Hotel Hyundai by Lahan Ulsan on the 13th. May 13, 2026 Yonhap News Agency. Photo by Yonhap News Agency

Kim Yongbum, policy chief, is attending the K-Chosun Future Vision Meeting presided over by President Lee Jae-myung at Hotel Hyundai by Lahan Ulsan on the 13th. May 13, 2026 Yonhap News Agency. Photo by Yonhap News Agency

View original image

In his Facebook post titled "Project Trinity: The Industrial Triad of the AI Era," Kim stated, "Although semiconductors, data centers, and physical AI may seem like separate industries at first glance, all three are required for AI to produce tangible results in the real world." He argued that for the value chain to be complete, computational infrastructure for training and inference, semiconductors capable of supporting this workload, and hardware to realize AI outputs in real environments must all be in place.


Kim assessed that with the global AI supply chain currently being reorganized, this moment presents a crucial opportunity for Korea. Previously, the supply chain worked such that the United States designed software and models, Taiwan manufactured advanced semiconductors, and China handled large-scale manufacturing. However, geopolitical risks, technological decoupling, and power shortages are now emerging as simultaneous challenges. Kim highlighted, "Korea is one of the few countries with advanced manufacturing capabilities, semiconductor technology, and robust power infrastructure all at once." He emphasized that if these three elements are combined, Korea can move beyond simply supplying parts and become a core hub supporting the entire AI supply chain.


Kim specifically noted that semiconductors and AI data centers are already closely intertwined. As AI models grow larger, the bandwidth between GPUs and memory becomes a key performance factor, and high bandwidth memory (HBM) has become increasingly important as the core technology for resolving this bottleneck. He explained, "HBM has become almost essential in AI accelerators, and Korean companies play a major role in its supply. The more AI infrastructure is built domestically, the greater the opportunity for Korean semiconductor companies to test and develop next-generation memory, packaging, and inference chips in real-world environments."


He identified power supply as the most critical variable for AI data centers. "Currently, the biggest bottleneck for AIDC investment is not money, but electricity," Kim said, noting that the ability to reliably supply gigawatt (GW)-scale power has become a matter of national competitiveness. He argued that it is advantageous for AI data centers to be located in non-capital regions, close to power generation facilities. If major consumers are established near generation sites, it can reduce the burden on transmission networks and serve as a catalyst for local investment in power generation and transmission infrastructure, thanks to assured demand.


Kim further described the construction of data centers as a process that creates an entirely new industrial ecosystem. Large AI data centers not only provide server space, but also foster related industries such as design, construction, operations, cooling, power management, networking equipment, and operational software. He pointed out, "Although data centers themselves do not require large numbers of on-site staff, the real value comes from the industries that develop around them, rather than employment within the facility." He added, "Establishing data centers outside the capital area helps decentralize industrial concentration and builds high-tech industrial bases in the regions."


Regarding the third pillar, physical AI, Kim predicted it could become a "second semiconductor" for Korea. He noted that although Korea was not a leading nation in semiconductors in the 1980s, it became a key player by leveraging focused investment and manufacturing capabilities before the supply chain was fully established. "Physical AI appears to be at a similar inflection point," he said. "If we enter the market after it is fully matured, we will only remain followers. To become a leader, we must step in as a core supplier while the supply chain is still being built."


Physical AI encompasses AI that operates in the physical world, including robotics, autonomous manufacturing, logistics automation, and assistive technologies. Kim observed that as labor shortages intensify in most advanced economies due to aging populations and a shrinking workforce, the structural demand for automation and robotics will inevitably continue to rise. He also noted that Korea's automotive factories, semiconductor lines, shipyards, logistics centers, and advanced manufacturing facilities provide an ideal environment for large-scale demonstration and development of physical AI.


The core of Kim's Trinity concept lies in not just developing the three industries individually, but linking them in a single virtuous cycle. Data centers train AI models, semiconductors efficiently run those models, and physical AI applies them in industrial settings. In turn, data generated by robots, factory operations, logistics, and facility sensors is fed back into data centers to further refine AI models. Kim referred to this as a "nationwide AI platform" and a "flywheel" (a rotational energy storage device).


Kim stressed, "Production capacity can be caught up, and technology can eventually converge, but a supply chain backed by accumulated data and operational experience cannot be easily replicated." He added, "Semiconductors are Korea's existing strength, AI data centers are an emerging infrastructure we can seize now, and physical AI is a future area we must secure before it's too late."



This proposal is being interpreted as a message that concretely articulates the Lee Jae-myung administration's AI transformation strategy from an industrial policy perspective. It reflects the view that as AI competition expands from model development into real-world foundations such as power, memory, data centers, manufacturing sites, and robotics, Korea should weave its existing industrial strengths into a cohesive national strategy. Kim concluded, "Now, as the global supply chain is being rebuilt, Korea has the opportunity to stand at its center. Project Trinity is a conceptual map to unite our scattered strengths into a single national strategy."


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing