SK Hynix will spend $38 billion building memory chip plants. Not upgrading existing plants. Not buying plants. Building new ones. From scratch. Because apparently the solution to high prices is flooding the market with supply in three to five years when nobody remembers why prices were high in the first place.
Memory prices surged. Demand soared. Supply couldn't keep up. Every semiconductor company on earth watched this happen and reached the same brilliant conclusion at the exact same time. Build more capacity. Right now. All of them. Together. Like a synchronized swimming routine performed by CFOs who failed economics.
Investors are watching closely for changes to the supply-demand imbalance. They found one. It's called $38 billion in new capacity that will come online precisely when every other chipmaker's new capacity comes online. The imbalance will reverse so hard it'll need physical therapy.
Retail traders saw this headline and bought semiconductor stocks. They read "demand soars" and stopped reading. They don't know what a capital cycle is. They don't know what marginal cost means. They think supply and demand is when line go up forever because people want chips.
SK Hynix executives presented this plan to the board with PowerPoint slides showing hockey stick revenue projections. Slide seventeen showed a footnote in six-point font that read "assumes competitors build nothing." Slide eighteen was approved budget allocation. Nobody scrolled back.
The memory chip business prints money until everyone remembers how to build memory chips. Then it prints apology letters to shareholders. SK Hynix just committed $38 billion to arriving fashionably late to both parties.
Photo by Brecht Corbeel on Unsplash
Leave a Comment