Samsung Shifts Over Half of 4nm Capacity to HBM4 Base Die
nashnova research
Samsung is allocating more than half of its 4nm foundry capacity to producing base dies for HBM4, accelerating AI-memory shipments — at the cost of squeezing capacity available to other foundry clients.
What is Samsung doing?
Per DigiTimes, Samsung Electronics is directing over half of its 4nm process capacity to base-die production for sixth-generation High Bandwidth Memory.
HBM4 — high-bandwidth memory designed for AI workloads — stacks memory chips on top of a base die built at an advanced node. That base die handles the control logic and data links.
This means → Samsung is betting its most advanced manufacturing resources on the "brain" of its memory chips.
Why make this bet?
The goal is clear: speed up shipments to major AI-chip customers.
Samsung has trailed SK Hynix in HBM for years. Tilting 4nm capacity heavily toward HBM4 is a play to close that gap before the AI-memory window narrows.
In plain terms = Samsung would rather sacrifice a share of its foundry revenue than fall further behind in AI memory.
What does this mean for other foundry clients?
With over half of 4nm capacity claimed by HBM4, the room left for external foundry customers shrinks.
This reflects a trade-off: Samsung chose internal HBM4 production over selling that same capacity to outside chip designers.
This means → companies that planned to tape out 4nm chips at Samsung now face a choice — wait for capacity to free up, or move orders to alternatives such as TSMC.
市场有风险，内容仅供研究参考，不构成投资建议。