Samsung Electronics to separate semiconductor foundry business into a new unit and lays out foundry process roadmap, revealing 4nm process generation
Ian King / Bloomberg :
Context & Ripple Effects
Samsung's move to carve its foundry operation into a separate unit reads as the opening gambit of its long campaign against TSMC rather than an accounting reshuffle: the same announcement commits to a public process roadmap that names a 4nm generation. The years of coverage that follow confirm the intent — Samsung's heavy investment in extreme ultraviolet lithography explicitly framed as a bid to rival TSMC, then a dedicated 5nm fabrication plant at Pyeongtaek built for contract chipmaking.
The roadmap discipline stuck: by 2018 Samsung claimed mass production of 7nm chips on EUV, and by late 2022 it had published a five-year plan reaching 2nm in 2025 and 1.4nm in 2027. The 2017 separation is the structural decision those commitments were hung on.
First-order effects
- Foundry customers get something they previously lacked at Samsung — a named unit with a published node cadence down to 4nm, making Samsung's contract capacity plannable alongside TSMC's instead of subordinate to Samsung's internal chip needs.
Second-order effects
- A standalone foundry competing for external orders puts direct pressure on TSMC's near-monopoly grip on leading-edge contract manufacturing, and as demand tightened Samsung later raised prices on advanced 4nm and 5nm orders by up to 15% — evidence the separated unit gained real pricing power.
Third-order effects
- If the separation-plus-roadmap pattern holds, contract chipmaking consolidates into a two-player race where credibility depends on publishing multi-year node plans and building captive capacity like Pyeongtaek — raising the capital bar that locks smaller foundries out of the leading edge.
The trend: Leading-edge foundry is consolidating around a Samsung-versus-TSMC duopoly in which organizational independence and published roadmaps are the currency of winning customer silicon.