[content] => 
    [params] => Array
            [0] => /forum/index.php?threads/samsung-uphill-battle-in-foundry-business.14691/

    [addOns] => Array
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2020771
            [XFI] => 1050170

    [wordpress] => /var/www/html

Samsung: Uphill battle in foundry business?

Daniel Nenni

Staff member
According to market research reports, the global wafer foundry market size was about US$82 billion in 2020, with TSMC having a 55% global market share, and Samsung taking only 15%. In terms of profit in the global wafer foundry sector, TSMC took nearly 85%, compared with less than 5% for Samsung. In other words, TSMC's profit in wafer foundry business was 17 times Samsung's. Evidently, the two players are competing in different divisions. But why has Samsung been so constantly provocative? Why is South Korea going all-out, vowing to be a dominant power in the global semiconductor industry by 2030?


Active member
Big name. Samsung desperately needs to improve its sales.

A tough nut because:
1. Samsung aims at leading edge
2. Not many companies can are even looking at leading edge besides usual CPU/GPU game, and will not do a leading edge chip even if the tapeout will come for free, as designing a chip even for regular DUV immersion finfet is much, much more expensive than what most generic ASIC fabless can afford
3. The few fabless clients who can do advanced logic at 45nm are held very tight by TSMC

Samsung really should've tried to take the market from bottom up. Their decision to chase margins in 201X at expense of building relationships is coming to haunt them.

Trying to ram your way into leading edge without extensive "platform" (oh this stupid word) is destined to fail for so many reasons, main being few will risk a botched leading edge tapeout.

On other hand, if you are already serving a sizeable generic, and legacy node clientele, you can have them upgrade to your newer nodes much more smoothly, and at a relaxed pace.
  • Like
Reactions: VCT