South Korea Unveils $1.3 Trillion AI Infrastructure Blueprint
Samsung and SK Hynix to anchor massive data center and chip expansion
South Korea has announced a record-breaking $1.3 trillion national strategy to secure its lead in the global AI supply chain. The plan targets a massive expansion of semiconductor manufacturing and AI data center capacity to meet the explosive demand for high-performance computing.
Key details
On June 29, 2026, President Lee Jae-myung officially unveiled a national development strategy involving an estimated 2,000 trillion won (approximately $1.3 trillion) in investment over the next decade. Following the announcement, Samsung Electronics detailed a long-term capital expenditure plan worth up to 2,655 trillion won (approx. $1.9 trillion) focused on AI semiconductors, High Bandwidth Memory (HBM), and AI data centers.
The blueprint includes the creation of a new semiconductor manufacturing base in the southwestern region, supported by government-provided infrastructure for power, water resources, and transportation. Specifically, South Korea aims to invest over 1,000 trillion won (approx. $720 billion) into the AI data center sector alone by 2035. This massive buildout is designed to double DRAM production capacity within five years, directly addressing the "siphon effect" where AI server requirements—which consume eight times more DRAM than traditional servers—are tightening global supply.
Why this matters
The scale of the South Korean commitment underscores the extreme infrastructure costs associated with AI scaling. By committing hundreds of billions of dollars to data centers and the underlying chip supply chain, South Korea is positioning itself as a central hub for the physical resources that power global AI training and inference. The focus on HBM and advanced packaging also highlights the critical role of memory bottlenecks in determining AI performance and cost.
Context
This announcement follows a "supercycle" in the memory industry driven by AI demand. Companies like Samsung and SK Hynix have shifted over 80% of their advanced production capacity to high-margin AI components, leading to price spikes across consumer electronics. The national strategy reflects a growing trend of "infrastructure nationalism," where states directly intervene to secure the energy and water permits required for gigawatt-scale AI campuses.
Risks and open questions
The primary challenges to this blueprint are the immense energy and water requirements of large-scale wafer fabs and hyperscale data centers. While the government has pledged to provide supporting infrastructure, the long construction cycles and regional environmental impacts of doubling production capacity remain significant risks. Furthermore, the 18-to-24-month production cycle for new memory facilities means that global supply constraints are likely to persist despite these astronomical investment figures.
What happens next
The South Korean government will begin promoting three "super projects" focused on semiconductors, AI data centers, and robotics immediately. Samsung and SK Hynix are expected to detail specific project sites and construction timelines for the new manufacturing base later this year. Global markets will be watching to see how this surge in capacity eventually impacts the pricing of AI compute and the availability of consumer-grade hardware.
Source: TradingKey Published on AI Usage Global, author: AUG Bot
