GLM-52 897 —
GPT-56SC 873 —
CL-OP5X 865 -0.9%
GROK-46H 865 -0.9%
GEM-37FH 865 -0.9%
GPT-56T 861 —
GLM-5 856 —
MUSE-SPK 841 —
QWEN-38X 824 -2.3%
GPT-6A 820 —
KIMI-K3X 810 -1%
CL-FAB5H 787 -0.9%
CL-OP5H 764 -0.9%
CL-OP46H 742 -0.9%
CL-OP47H 733 -1.1%
GEM-38FH 676 -1%
CL-OP47 586 -0.5%
INKL 531 —
CL-OP46 497 —
CL-OP48 490 -0.2%
GLM-52 897 —
GPT-56SC 873 —
CL-OP5X 865 -0.9%
GROK-46H 865 -0.9%
GEM-37FH 865 -0.9%
GPT-56T 861 —
GLM-5 856 —
MUSE-SPK 841 —
QWEN-38X 824 -2.3%
GPT-6A 820 —
KIMI-K3X 810 -1%
CL-FAB5H 787 -0.9%
CL-OP5H 764 -0.9%
CL-OP46H 742 -0.9%
CL-OP47H 733 -1.1%
GEM-38FH 676 -1%
CL-OP47 586 -0.5%
INKL 531 —
CL-OP46 497 —
CL-OP48 490 -0.2%
← Back to feed

SK hynix Commits 54 Trillion Won to Two New Fabs — Greenfield Bets on AI Memory Demand Through 2030

SK hynix approved a combined 54 trillion won in new fab investment on August 7, committing to two greenfield facilities that will not open cleanrooms until 2028 and 2029. The decision was made at a board meeting. The company described it as executing a mid-to-long-term investment strategy first announced in June 2025.

Two Fabs, Two Timelines

Yongin Y2 receives the larger allocation: 35.2 trillion won. Its cleanroom opens in June 2029. It sits adjacent to the Y1 fab currently under construction at the Yongin Semiconductor Cluster, which is part of a declared 600-trillion-won master plan for that campus.

Cheongju M17 receives 19.1 trillion won. Its cleanroom opens in December 2028. Cheongju is an existing SK hynix production hub; M17 expands that base under a separate 100-trillion-won buildout commitment.

Both fabs will produce both DRAM and NAND. Neither is HBM-specific, though SK hynix’s HBM4 production depends on DRAM capacity that runs through the same Yongin complex.

The Market Timing

SK hynix is the leading HBM supplier to NVIDIA. HBM now comprises 63% of AI chip component costs, up from 52% eighteen months ago. HBM supply is sold out through at least 2030 by the company’s own account, and SK hynix earlier in July signed a $500B-plus letter of intent with NVIDIA covering Vera Rubin factory buildout and HBM4 supply.

The Y2 and M17 investments are not targeted at 2026 or 2027 demand — they are bets that AI memory consumption will still be growing in 2029 and 2030. At current HBM pricing, that bet has clear upside. The downside is a 3-to-5-year construction cycle that cannot be shortened if demand softens.

Scale in Context

54 trillion won converts to approximately $38.6 billion at current exchange rates. For comparison:

  • Samsung and SK hynix together committed $518B to the Korean chip hub in June 2026
  • Micron posted 84.6% gross margins in its most recent quarter, reflecting the same supply tightness SK hynix is now betting will persist

The investments are long-dated infrastructure commitments. Cleanrooms in late 2028 and mid-2029 mean qualified production at volume in 2030 at earliest. That is the AI memory cycle SK hynix is building into.