Samsung and SK Hynix Plan $646B South Korea Chip Expansion

The numbers are striking enough to double-check: Samsung and SK Hynix are reportedly preparing to announce a combined investment package approaching 1,000 trillion won, or roughly $646 billion, over the next ten years, according to Bloomberg. If the figures hold, this would represent the largest corporate investment commitment in South Korean history, a sum that reportedly exceeds South Korea's own annual government budget of 728 trillion won and approaches half of the country's annual gross domestic product.

The plan, reportedly set to be unveiled formally on Monday, centers on building out a second major chip cluster in South Korea, with a focus on expanding the country's industrial base beyond the Seoul metropolitan area. The spending reportedly spans semiconductors, artificial intelligence, next-generation batteries, and advanced displays. Samsung's P5 facility in Pyeongtaek is expected to be operational by 2028; SK Hynix's M15X facility is slated for mid-2027.

The strategic logic is not hard to find: Samsung and SK Hynix together hold a dominant share of the global HBM market, the segment of memory chips specifically designed for the massive parallel processing that AI models require. As AI training and inference workloads have scaled, HBM supply has become a recognized chokepoint in broader AI infrastructure buildouts. Capital commitments at this scale are a bet that demand continues to hold, and that securing production capacity now justifies the risk.

The honest caveat is that the 1,000 trillion won figure is described as the upper-end projection, not a signed commitment, and a ten-year spending package will vary considerably with memory market cycles, which both companies have navigated before. What the reporting does not give you is how the total breaks down between Samsung and SK Hynix individually, or how much is specifically earmarked for HBM versus batteries, displays, and other segments.

If the announcement lands as reported, the clearest near-term beneficiaries are the semiconductor equipment and materials suppliers feeding new fab construction, and the AI infrastructure operators who have been managing memory scarcity as a hard constraint on scaling.