WYP
News Guru
Samsung wants to be the top foundry by 2030.
Read more about Samsung's plans to invest $115 billion in Semiconductors.
Read more about Samsung's plans to invest $115 billion in Semiconductors.
Last edited:
Samsung plans to invest $115 billion to become the leaders the semiconductor industry
- Thread starter WYP
- Start date
WYP
News Guru
Remember that TSMC is the top dog right now, Samsung investing will help make them more competitive with TSMC. Being more competitive with TSMC is good news, as it will give product makers more choice when it comes to leading-edge nodes and increases the high-end chip-making capacity that the market has available to it.
Ideally we would have more than just TSMC and Samsung competing, but at least TSMC and Samsung are competing, and are committed to the foundry market for the next decade.
Samsung committing this money is good news. The fact that companies are leaving the leading-edge market with each big transition is bad news, but nothing can really be done about that. The foundry market needs an insane amount of money to operate, and GlobalFoundries didn't have the investment they needed to reach for the top anymore.
Semiconductor technology takes at 2-3 decades to go from research to practical examples, even after practical prototypes it's often a decade or so till we have techniques to manufacture the technology en masse, so since the hardest bit really is going from practical prototypes to finding ways of mass production you never really get surprises in this industry.
Silicons practical uses were discovered around the 50's but it wasn't till the 70's we could really make use of it, FinFETs prototypes came in 2001 but ofc it wasn't until 2013 we had shipping products, EUV R&D started in the 1990's, with the first prototypes coming around 2006, first practical shipping products this year.
It looks like the next big thing will be transition metal dichalcogenide monolayers (2D exfoiliatable atom layer materials of metalic atoms), similar in structure to graphene but the lack of a band gap makes graphene pretty impractical, however there's many hundreds of these materials currently in research and the most promising atm to replace silicon looks like indium selenide, which is still a decade or so away from practical uses, and maybe one of the hundreds of other "hot" materials in this space like molybdenum diselenide maybe in particular will turn out to be more practical.
Optronics could be somewhat disruptive but the use cases where it's most beneficial don't seem wide spread, same with things like quantum computing, great for some tasks but abysmal for others.
Basically, silicon still has a very long time left, we're going to be squeezing whatever we can out of it over the next decade without a doubt. But to be honest, fabs don't have to be cutting edge to be useful, most nodes are used for a decade or so after they're cutting edge with slight retooling and updates for cheaper mass manufacturing, the bigger risks are more along the lines of natural disasters.
Silicons practical uses were discovered around the 50's but it wasn't till the 70's we could really make use of it, FinFETs prototypes came in 2001 but ofc it wasn't until 2013 we had shipping products, EUV R&D started in the 1990's, with the first prototypes coming around 2006, first practical shipping products this year.
It looks like the next big thing will be transition metal dichalcogenide monolayers (2D exfoiliatable atom layer materials of metalic atoms), similar in structure to graphene but the lack of a band gap makes graphene pretty impractical, however there's many hundreds of these materials currently in research and the most promising atm to replace silicon looks like indium selenide, which is still a decade or so away from practical uses, and maybe one of the hundreds of other "hot" materials in this space like molybdenum diselenide maybe in particular will turn out to be more practical.
Optronics could be somewhat disruptive but the use cases where it's most beneficial don't seem wide spread, same with things like quantum computing, great for some tasks but abysmal for others.
Basically, silicon still has a very long time left, we're going to be squeezing whatever we can out of it over the next decade without a doubt. But to be honest, fabs don't have to be cutting edge to be useful, most nodes are used for a decade or so after they're cutting edge with slight retooling and updates for cheaper mass manufacturing, the bigger risks are more along the lines of natural disasters.
Last edited:
Luckily pretty much all of the research on this stuff is in academia rather than commercial companies, usually there's only commercial interests after the science behind it has been fully understood through research which I think is why we usually get a 10 year lead, that's roughly when it goes from open academic research projects to more secretive commercial ventures looking into viable manufacturing processes & similar.
Similar threads
