Samsung’s president and head of semiconductor business, Kyung Kye-hyun, has stated that the company aims to surpass TSMC’s microchip fabrication technology within five years. He has acknowledged that Samsung’s current foundry technology is one or two years behind TSMC’s. However, Samsung believes that it has an advantage over TSMC as the industry moves towards 2nm technology.
Samsung claims that its gate-all-around (GAA) transistors for its current 3nm node deliver superior performance, lower power consumption, and a reduction in chip area compared to its previous node. TSMC has retained older FinFET transistor designs in its new 3nm node and will move to GAA transistors only when it transitions to 2nm. Samsung believes that it has an advantage when it comes to implementing GAA transistors, which could enable it to overtake TSMC in microchip fabrication technology.
The Inconsistency in Samsung’s Narrative
Industry analysts suggest that Samsung’s 3nm technology is comparable to TSMC’s older 5nm node. TSMC’s 5nm node hit the market two years before Samsung’s 3nm technology, giving TSMC a two-year advantage. Samsung seems to be drawing an equivalency between its upcoming 2nm technology and TSMC’s 2nm node, despite being two years behind TSMC. It is unclear how Samsung’s 2nm node will suddenly surpass TSMC’s technology in a single step.
Intel’s recent claims for regaining leadership over TSMC seem more feasible than Samsung’s claim. Intel is promising four new nodes in five years, which could enable it to overtake TSMC if it delivers on its timeline.
Healthy competition between TSMC, Samsung, and Intel would be advantageous for ever-faster graphics and processors in the future. Currently, AMD makes its latest GPUs on TSMC’s 5nm technology, while Nvidia uses TSMC’s 4nm, which is a refined version of 5nm. Meanwhile, Intel makes CPUs on its own Intel 7 process, and AMD’s latest processors use 5nm TSMC chiplets. If the big players in chip production manage to up their game, it could lead to tougher competition and better value CPUs and GPUs.