Korea steps up its AI industry push with “three mega projects” while simultaneously rolling out price stabilization measures

Overview

This report discusses the Korean government's newly announced “three mega projects” and price stabilization measures. The industrial policy section focuses on three pillars—semiconductors, Physical AI, and AI data centers—with the goal of expanding Korea's manufacturing, chip, and computing-power base in the AI era through long-term public-private investment and infrastructure support. The price policy section aims to drive inflation lower through fiscal subsidies, lower fuel price caps, food discounts, tariff cuts, and freezes on public service prices.

Core views

Goldman Sachs's core view is that the Korean government is stepping up support for both the AI industry chain and price stability. Combining government projects with long-term corporate capital expenditure plans from Samsung, SK Group, and others, Korea's AI-related investment could reach an average of KRW380 trillion per year over the next decade, about 10% of GDP, of which semiconductor-related investment would account for about 6% of GDP. This implies sustained policy and capital support for Korea's semiconductor, AI manufacturing, data center, power and cooling, and NPU industry chain segments. Meanwhile, the price stabilization measures cover fuel, food, utilities, transportation, and other items, and are expected to be mainly reflected in the July CPI.

Analysis framework

The report interprets policy announcements and aggregates capital expenditure, combining the Korean government's funding commitments, capacity targets, and infrastructure arrangements under the “three mega projects” with the long-term investment plans disclosed by companies such as Samsung, SK Hynix, and SK Telecom, and further analyzes the coverage of price stabilization measures within the CPI basket and their potential timing impact.

Methodology notes

• Industrial policy analysisThree mega projects framework

  Three pillars: semiconductors, Physical AI, and AI data centers

  This framework is used to break down the main directions, targets, and funding commitments of the Korean government's long-term AI support policies and to identify the relevant beneficiary industry chains.

• Capital expenditure estimationAggregation of government and corporate long-term investment plans

  Estimate of AI-related investment as a share of GDP

  The report combines government projects with the long-term investment plans of Samsung, SK Hynix, SK Telecom, and other groups to estimate that AI-related investment over the next decade could average about 10% of GDP annually, of which semiconductors would account for about 6%.

• Inflation policy assessmentCPI basket coverage analysis

  Coverage and transmission of price stabilization measures to CPI

  The report assesses policy coverage across items such as fuel, food, utilities, transportation, and subsidies for low-income households, and judges that most of the relevant measures will be reflected in the July CPI.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Korean semiconductor industry chain

  Direct beneficiary

  Strengths

  Receives support for large-scale capacity expansion, advanced packaging, emerging chip markets, and a presidentially led special committee.

  Weaknesses

  Projects have long cycles and high capital intensity, and some investment timelines have not yet been clarified.

  Comparison

  Compared with typical cyclical semiconductor investment, this plan places more emphasis on strategic capacity for the AI era, next-generation memory, and regional manufacturing base development.

  Risks

  Execution delays, power and water resource constraints, demand-cycle volatility, and intensifying global competition.

• AI data centers, NPU, power, and cooling ecosystem

  Industry chain beneficiary

  Strengths

  The KRW550 trillion initial investment and 8.4GW capacity target provide demand support for data center infrastructure and supporting technologies.

  Weaknesses

  Data center construction is highly dependent on power, land, approvals, and cooling resources.

  Comparison

  Compared with standalone cloud-computing expansion, this plan places greater emphasis on national-scale computing infrastructure and a domestic NPU ecosystem.

  Risks

  Power costs, grid connection progress, technology substitution, overseas investment, and uncertainty around funding execution.

• Physical AI and robotics

  Policy-supported development

  Strengths

  The government's goal is to make Korea one of the world's top three AI robotics powers and achieve global leadership in Physical AI by 2030.

  Weaknesses

  Foundation models, data, mass-production capability, and commercialization still need to be validated.

  Comparison

  This direction extends Korea's traditional manufacturing strengths into AI robotics and manufacturing intelligence.

  Risks

  Insufficient technological maturity, slow enterprise adoption, global competition, and talent bottlenecks.

• Korean inflation and consumption environment

  Short-term buffer

  Strengths

  Support for fuel, food, utilities, and transportation helps contain short-term CPI pressure.

  Weaknesses

  Some measures are price caps or subsidies, and their sustainability depends on the fiscal and energy price environment.

  Comparison

  Unlike long-term industrial investment, price stabilization policy is more of a short-term macro management tool.

  Risks

  Oil price rebound, food price volatility, fiscal burden, and re-inflation risk after the withdrawal of price freezes.

Key data

• Total government semiconductor projectsAbove KRW881 trillionIncludes KRW800 trillion in the southwestern region, KRW81 trillion in the Chungcheong region, and support for a materials, parts, and equipment innovation hub in the southeastern region.
• Support for emerging chip marketsKRW30 trillion, over the next 15 yearsFor high-growth markets such as next-generation memory, edge AI, and defense chips.
• AI manufacturing investmentKRW20 trillion, by 2030Aims to drive AI transformation in manufacturing and create KRW100 trillion in added value.
• Initial AI data center investmentKRW550 trillion, 8.4GWMajor groups such as SK Group, GS Group, and Naver will build world-class AI data center capabilities.
• Long-term data center capacity18.4GW, by 2035The report says SK Group will further expand capacity by 10GW, lifting total capacity.
• Samsung domestic long-term investmentKRW2,450 trillion, by 2040Of this, about KRW2,100 trillion is for semiconductors.
• SK Hynix long-term semiconductor investmentKRW1,100 trillionOf this, KRW600 trillion is for the Yongin semiconductor cluster project.
• AI-related investment intensityAbout KRW380 trillion/year, about 10% of GDPEstimated average over the next decade; of this, semiconductor-related investment accounts for about 6% of GDP.
• Fiscal spending for price stabilizationKRW1 trillionIncludes agricultural product subsidies, discount campaigns, tariff reductions, transportation support, and assistance for low-income households.
• Fuel price cap adjustmentKRW150 per liter reduction, about 10%Corresponds to a lower local fuel price cap following the recent decline in global oil prices.
• Food discount fundingAbout KRW400 billionFor major discount campaigns in July and August, covering eggs, rice, and certain seafood products.
• CPI basket coverageGoods-related measures cover about 12% of the CPI basket, and public service freezes cover about 15%By subcategory, goods basket coverage is about 30%, and services basket coverage is about 26%.

Impact & implications

This policy mix has positive long-term implications for Korea's AI and semiconductor industry chain: on one hand, large-scale investment in capacity, packaging, data centers, and supporting power and water resources will strengthen Korea's strategic position in AI hardware and computing infrastructure; on the other hand, the government's price stabilization measures will help cool short-term inflation, providing some buffer for the policy environment and household consumption. From an investment perspective, key areas to watch include semiconductor capital expenditure, AI data center construction, power and cooling solutions, the NPU ecosystem, robotics, and AI transformation in manufacturing.

Risks

• The scale of long-term government and corporate investment is enormous, and there is uncertainty around the actual pace of implementation, funding sources, and project returns.
• Semiconductor and AI data center projects depend on stable power, water resources, land approvals, and infrastructure construction, creating high execution risk.
• Technology paths related to AI chips, NPU, robotics, and Physical AI are still evolving rapidly, with risks of technological substitution and competition.
• Global semiconductor cycles, AI demand, trade policy, and geopolitics may affect investment efficiency and capacity utilization.
• Price stabilization measures may suppress CPI in the short term, but if energy or food prices rebound, policy effectiveness may weaken.
• Utility price freezes and subsidy policies may increase the burden on public finances or state-owned enterprises, and the pace of subsequent withdrawal needs to be monitored.

What to watch

• Follow-up budget arrangements, project timelines, and specific implementing entities for Korea's “three mega projects”.
• Approval, construction commencement, and capacity build-out progress of semiconductor bases in Yongin, southwestern Korea, the Chungcheong region, and the southeastern region.
• Actual capital expenditure delivery under the long-term investment plans of Samsung, SK Hynix, SK Telecom, SK Group, GS Group, and Naver.
• Progress in AI data center capacity construction, power supply, cooling solutions, the NPU ecosystem, and renewable energy expansion.
• Korea's July and subsequent CPI data, especially subcomponents for fuel, food, and public service prices.
• Whether measures such as lower price caps, tariff cuts, and utility price freezes are extended, and their impact on inflation upon exit.