Samsung to Shut Down All 8-Inch Wafer Foundry Lines by 2030
nashnova research
Samsung Electronics has told customers it will close every 8-inch foundry line by end-2030 — its first firm deadline. Korean alternatives are running near full capacity, leaving fabless chip firms facing steep re-qualification costs and a tightening supply chain.
What exactly is Samsung shutting down?
Samsung has notified clients that all 8-inch wafer foundry lines — covering 65 nm, 70 nm, 90 nm, 0.13 µm and 0.18 µm process nodes — will close by end-2030.
This means → Samsung is not trimming capacity; it is exiting 8-inch contract manufacturing entirely.
The company has already stopped accepting new 8-inch product orders this year; the timeline is accelerating.
Who gets hit hardest?
The most exposed are Korean fabless chip designers — Imagis, ABOV Semiconductor, Dongwoon Anatech, Silicon Mitus and LX Semicon — which rely on Samsung's mature lines for display drivers, power-management ICs, microcontrollers and image sensors.
In plain terms = these firms own no fabs; Samsung's legacy lines *are* their factory. Once those lines close, they must find a new one.
Analog and high-voltage chips are especially hard to move — migrating a design means layout changes, new masks, full reliability testing and customer re-qualification.
Where can orders go — and why is it so difficult?
Korea has only two other major 8-inch foundries: DB HiTek and SK Keyfoundry. Both are nearly full — DB HiTek's two fabs ran at an average 94.9% utilization in H1 2026, shipping roughly 154 000 wafers per month.
This means → spare capacity for Samsung's migrating orders is extremely limited. DB HiTek has already raised prices by up to 30% for some clients this year, with Chinese customers seeing the steepest hikes.
Taiwan is similarly tight: UMC hit 92% utilization in Q3, with some 12-inch fabs at full load; PSMC raised prices 10–15% for clients without annual contracts and is negotiating a further 10–20% increase for 2027.
Can Chinese foundries absorb the overflow?
Mainland Chinese fabs are a theoretical alternative for some products, but export controls limit what can actually be transferred.
In plain terms = some chips could move to Chinese lines, but the regulatory fence blocks a complete solution.
Why is Samsung doing this — and where do the resources go?
Samsung plans to redirect headcount and equipment investment from 8-inch operations into 12-inch advanced nodes, betting heavily on 4 nm and 2 nm.
Its latest quarterly results show the foundry division improving on the back of rising HBM base-die sales, more advanced-node orders, and new 2 nm HPC wins from a major hyperscaler.
This reflects a clear strategic call: mature-node foundry work is low-margin and fiercely competitive — Samsung would rather concentrate resources on advanced nodes to compete for high-end clients.
Can the supply gap be closed before 2030?
Samsung executive Mi-jung Noh said in September that high-performance computing's share of foundry revenue has risen from 5% in 2017 to 28%, with a target of 66% by 2029 — the advanced-node pivot is accelerating.
Samsung plans to begin third-generation 2 nm mass production in 2028 and move to 1.4 nm in 2029; its new Taylor, Texas fab is expected to come online next year.
Global installed 8-inch capacity is forecast to grow 3% this year to 7.9 million wafers per month, according to SEMI — but that figure includes IDM captive capacity, and the share actually available to external customers is hard to isolate. Whether alternative capacity expands fast enough before Samsung's 2030 deadline is the central supply-chain question for every affected chip firm.
市场有风险，内容仅供研究参考，不构成投资建议。