1. the Great Migration of Capital and the Dawn of the 'Physical AI' Era
the South Korean economy is passing through a huge inflection point. the astronomical investment plans announced by the four major Korean conglomerates - Hyundai Motor, Samsung, SK, and LG - totaling nearly 800 trillion won ($810 billion) herald a "qualitative transformation" that will fundamentally change the DNA of the industry, rather than simply expanding facilities. whereas past investments have been focused on being a fast follower, catching up to the technology of the developed world and gaining mass production capabilities, the current investments clearly indicate a shift to a "first mover" strategy to preempt the technological Singularity.
the key words running through this massive flow of capital are "Physical AI" and "Infrastructure". if generative AI, represented by ChatGPT, led the revolution in the digital world, it is now coming down to the physical world by acquiring a body in the form of a robot. essential to this process are 'foundries' to produce robots, 'semiconductors' to power AI, and 'energy and data centers' to support it all.
this report analyzes the details of the 800 trillion won investment in detail and draws insights into the investment and high-value keywords to watch for on major blog platforms such as Naver and T-story. in particular, we delve deep into the causal relationships between Hyundai's robot foundry strategy, Samsung and SK's semiconductor cluster war, and the Haenam Solar City project in Jeollanam-do, which will solve their Achilles' heel: electricity.
2. hyundai Motor Group: The end of mobility and the birth of the 'robot foundry'
hyundai Motor Group's 125 trillion won investment is a declaration that it will destroy itself as an automaker and reinvent itself as a provider of intelligent mobility solutions. at its center is an innovative business model called 'Robot Foundry'.
2.1 Robot Foundry: Manufacturing as a Service (MaaS) and Economies of Scale
hyundai Motor Group's 'Robot Foundry' is a concept that transplants the successful methods of the semiconductor industry to the robot industry. just as semiconductor fabless (design) companies entrust production to foundries (outsourced production), Hyundai aims to create an ecosystem where robot service companies can focus on software and service development by standardizing the hardware platform of robots and establishing a mass production system.
this is a strategy to overcome the limitations of 'low-volume production of various types of products' that have been a problem in the robot industry. to this end, Hyundai has established a three-stage robotics plan that follows the "Make, Test, and Teach" path. in the 'Teach' phase, Hyundai uses thousands of Nvidia's AI chips to perform virtual and real-world simulations to dramatically improve the robot's intelligence.
phasekey Takeawaysstrategic Goal Make build a robotics foundry standardize hardware and reduce mass production costs Test test virtual/reality integration ensure responsiveness to different environmental variables Teach aI learning at scale cultivate autonomous judgment based on physical AI
2.2 Physical AI: A data revolution beyond LLMs
the technical core of Hyundai's investment is physical AI. unlike traditional generative AI, which generates text or images, physical AI is AI that understands, predicts, and acts on the physical laws of the real world.
according to Samsung Securities' analysis, physical AI outperforms LLM in terms of the scale of input and output data. it generates about 10,000 times more tokens (data processing units) than LLM for autonomous driving training and 100,000 times more for robotics training. this is because robots require massive amounts of sensor data and physics math information from walking, picking up objects, and balancing in the real world.
this is why Hyundai Motor Group has preemptively secured 50,000 copies of Nvidia's next-generation GPU Blackwell and plans to build a 6 trillion won AI data center in Korea. this is not just an expansion of IT infrastructure, but an essential investment to complete the "Fleet Data Loop" to train the brains of physical AI.
2.3 MobED and Atlas: at the forefront of commercialization
this strategy is already materializing into concrete products. mobED (Mobile Eccentric Droid) from Hyundai Motor's Robotics Lab and Atlas from Boston Dynamics are two examples.
MobED (mass-produced model): it features an "eccentric mechanism" with four independently driven wheels, allowing it to level out and navigate stairs or unevenness up to 20 centimeters high. it can be used as a platform for various service robots such as delivery, guidance, and filming, and the company aims to sell 10,000 units in the next three years.
atlas (industrial application): atlas, a humanoid robot, has begun pilot operations at the Hyundai Motor Group MetaPlant America (HMGMA) plant in Georgia, USA. atlas demonstrated its advanced physical AI ability to recognize and react to situations on its own, even when a human operator closes a part box or interferes. this has been likened to an "iPhone moment" for the robotics industry, as it marks the first time a robot has been taken out of a controlled lab and into a real-world industrial environment with unpredictable variables.
3. the semiconductor supercycle: SK and Samsung's 600 trillion won game
the semiconductor industry, a pillar of South Korea's exports, is riding the AI tidal wave and making "second startup" level investments. SK hynix and Samsung Electronics are battling to maintain their supremacy in memory semiconductors while expanding their dominance in the system semiconductor and foundry markets.
3.1 SK Group: HBM Monopoly and Yongin Semiconductor Cluster
SK hynix has established unrivaled leadership in the high-bandwidth memory (HBM) market, which addresses a key bottleneck in AI computing. to consolidate this position, the company is expected to invest KRW 128 trillion by 2028, with about KRW 600 trillion going to the Yongin semiconductor cluster alone in the long term.
structural changes in investment: Due to the miniaturization of semiconductor processes and the explosion of AI demand, input costs are increasing exponentially. the investment cost per 10,000 square meters of clean room has nearly tripled from KRW 7.5 trillion in 2019 to about KRW 20 trillion in 2025 based on Cheongju M15X. this shows the harsh reality of "no money, no technology".
the need for deregulation: SK hynix is strongly calling for the deregulation of holding companies in order to finance its massive investments. under the current Fair Trade Act, a holding company must hold 100% of the shares of a subsidiary company (SK hynix) to establish a new subsidiary, which prevents it from attracting outside capital and is a barrier to investment. SK argues that if this regulation is relaxed, a special purpose corporation (SPC) could be used to spread the financial burden and preserve the "golden time" for investment.
3.2 Samsung Electronics: KRW 450 trillion investment and 'turn-key' strategy
samsung Electronics is investing KRW 450 trillion in Korea alone over the next five years to maximize its strengths as the only integrated device manufacturer (IDM) that spans memory, foundry, and packaging.
samsung's core strategy is 'turn-key' services. By providing AI chip fabless companies with design support, foundry production, memory supply, and advanced packaging, Samsung offers a one-stop solution to shorten delivery times and optimize chip performance. this is a strong incentive for global big tech companies such as Nvidia and AMD to consider Samsung as a partner.
4. infrastructure paradox: Power shortages and data center relocation to rural areas
behind the brilliant blueprint of the 800 trillion won investment, there are two critical risks: power and water. AI data centers and advanced semiconductor factories are "electricity hippos" and "water hippos," and the infrastructure of the Seoul metropolitan area is already saturated.
4.1 The collapse of the metropolitan power grid and the rise of distributed energy
due to the concentration of data centers in the Seoul metropolitan area, permits for new data centers have virtually stopped, and KEPCO's expansion of the power grid is facing opposition from residents. this is in line with the trend of global big tech companies directly contracting with nuclear power plants to secure power or developing their own small modular reactors (SMRs).
in the U.S., Google, Amazon, and other big tech companies have spent more than $140 billion on mergers and acquisitions (M&A) of power companies, proving that power is the hegemony of AI. in Korea, "power security" has emerged as a key variable for industrial competitiveness, and power-related industries known as "Itaewon" (secondary battery, solar, nuclear power) are emerging as new leaders.
4.2 'Solar City' in Haenam, Jeollanam-do: A mecca for RE100 data centers
in the midst of this crisis, the 'Solarseado ' enterprise city of Haenam, Jeollanam-do is emerging as a unique alternative. it has the only large-scale RE100 (100% renewable energy) site in Korea.
overwhelming renewable energy potential: Built on reclaimed land in Haenam-gu, Solarseado is nine times the size of Yeouido in Seoul and has the potential to generate 5.4 GW of solar power. that's the equivalent of five nuclear power plants.
data center park: South Korean companies such as Samsung C&T, LG CNS, and SK, as well as global asset managers such as BlackRock, plan to invest up to KRW 8 trillion to build a 1GW data center park.
national AI Computing Center: The National AI Computing Center, led by a consortium of Samsung Electronics and Samsung SDS, is likely to be located in Haenam. the center will have 50,000 Nvidia GPUs by 2030 and will serve as a hub for AI research in Korea.17
haenam Solar City has great potential to grow into Asia's data center hub, as it is optimally located to solve the power shortage in the metropolitan area and meet the RE100 needs of global big tech companies.
5. the Small Business Revolt: Finding the Hidden Beneficiaries
large-scale investment by large companies inevitably leads to a trickle-down effect for smaller companies in the value chain. in this investment cycle, the areas to watch are robotic gearboxes, liquid cooling, and semiconductor substrates.
5.1 Localization of robot decelerators: breaking the Japanese monopoly
precision gearboxes, which account for more than 30% of the cost of a robot, have long been a market dominated by Japan's Harmonic Drive Systems. However, for Hyundai's robot foundry strategy to succeed, it is essential to reduce costs by localizing components.
sPG (SPG): sPG has successfully mass-produced precision reducers (SH, SR) for robots and is supplying them to Rainbow Robotics. it has recently established a mass production system and dramatically shortened the delivery time from one year to one month, which is a strong competitive advantage that can immediately respond to the rapidly changing robot market demand.
SBB Tech: Localized the first harmonic type reducer in Korea and is expanding its application to semiconductor transfer robots and defense.
5.2 Immersion Cooling: cool down the heat of the data center
As the heat generation of AI chips has skyrocketed, conventional air cooling has reached its limits. this is why immersion cooling, which cools servers by immersing them directly in a specialized solution that doesn't conduct electricity, is becoming a must-have technology.
related Stock Trends: kNSOL is partnering with Submer, the world's No. 1 liquid cooling company, while GST and Samsung Air Conditioning are also rushing to enter the market with their own technologies. in addition to dramatically reducing power consumption, liquid cooling can also help solve water scarcity by reducing the amount of cooling water used.
5.3 LG Innotek and LG Chem: Advancing Materials
LG Group is investing KRW 100 trillion to strengthen its competitiveness in materials and components.
LG Innotek (FC-BGA): lG Innotek is expanding its business from camera modules to FC-BGA, an AI semiconductor substrate. it has invested heavily in its Gumi plant to secure production capacity and is expected to directly benefit from the growing demand for AI servers.
LG Chem (EUV photoresist): after the Japanese export restrictions, LG Chem successfully localized EUV photoresist and is transforming into a semiconductor materials company. the company is also expanding its cathode production capacity to 320,000 tons by 2027 at its Cheongju and Gumi plants to strengthen its dominance in the battery materials market.
6. conclusion and Investment Insights: Roadmap to 'AI G3'
the KRW 800 trillion investment is the final push for Korea to become more than just a semiconductor manufacturing powerhouse, but an "AI G3" (US, China, and Korea). the success of this investment depends on the harmonious combination of hardware (semiconductors, robots) and infrastructure (power, data centers).
here are the key points for investors to pay attention to
physical AI value chain: Pay attention to reducer (SPG, SBB Tech) and sensor component companies that can be integrated into Hyundai's robot foundry ecosystem.
power infrastructure supercycle: AI data center expansion inevitably leads to a surge in demand for power facilities. look for long-term interest in wires (LS Cable), transformers (HD Hyundai Electric), and distributed energy stocks.13
evolution of the data center: Don't miss the trend shift in cooling technology from air cooling to water cooling to liquid cooling. liquid cooling stocks (KNSOL, GST) are more than just a theme, they are outperformers.
rise of regional hubs: Regulations and power shortages in the metropolitan areas will make regional data center clusters like Haenam Solar City a new investment destination.
in the end, "whoever controls data controls AI, and whoever controls power controls the future." the winners in the coming era of physical AI will be those who are first to capitalize on the $800 trillion flow of money.
appendix: Summary of key investment areas and related stocks
categorykey Issueskey Related Stocks/Keywordsinvestment Point robotics robot foundry, Physical AI hyundai Motor, SPG, Rainbow Robotics reducer localization, mass production system semiconductor HBM, Turnkey Service SK Hynix, Samsung Electronics, Hanmi Semiconductor HBM3E exclusive, advanced packaging infrastructure power shortage, data center LS Cable & System, HD Hyundai Electric, Hyosung Heavy Industries power grid aging replacement, AI power demand cooling liquid cooling, Thermal management kNSOL, GST, Samsung Air Conditioning essential technologies for high heat GPU cooling materials semiconductor substrates, battery materials LG Innotek (FC-BGA), LG Chem Benefit from AI chip performance advancement