The Stack
High Bandwidth Memory as the scarce stack next to AI GPUs — who supplies it, why allocation binds, and where to read next.
HBM
Samsung Electronics gets paid on the memory line, which the company put at KRW 140.7 trillion for the quarter to 30 September 2026, while System LSI and the foundry remain inside a KRW 6.4 trillion remainder that was the same size a year earlier.
HBM
Micron designs and makes DRAM — including HBM stacks — and NAND: it gets paid when bits ship, while $12.75 billion of customer deposits lock future supply and Idaho wafers still wait until mid-2027.
HBM
SK hynix builds HBM stacks, sells them under long agreements to a short customer list, and gets paid when those stacks ship — allocation is the stuck step.
Home / The Stack / Semiconductors / HBM
High Bandwidth Memory sits next to the GPU. Without enough HBM stacks, finished AI accelerators do not ship — even when wafer starts look fine.
What it is: Stacked DRAM that feeds bandwidth into AI GPUs and accelerators.
Who does it: Memory makers supply HBM; GPU designers and foundries consume it under allocation.
Where it binds: Advanced HBM capacity and customer qualification — not commodity DRAM.
Read next: HBM, Explained · back to Semiconductors.
This hub collects HBM-tagged articles. It is a map of the memory gate inside the semiconductor layer — not a buy list.
