South Korea Mulls Creating 'KSMC' Contract Chipmaker To Compete With TSMC (tomshardware.com) 11
An anonymous reader quotes a report from Tom's Hardware: Although Samsung Foundry is a major chip contract manufacturer, the South Korean government mulls creating a government-funded contract chipmaker tentatively called Korea Semiconductor Manufacturing Company, KSMC, reports The Korea Biz Wire. Industry experts and academics have proposed the initiative.
The Semiconductor Industry Association's Ahn Ki-hyun called for a long-term government investment. Experts project that an investment of KRW 20 trillion ($13.9 billion) in KSMC could result in economic gains of KRW 300 trillion ($208.7 billion) by 2045. However, the big question is whether $13.9 billion is enough to establish a chipmaker. Another concern about publicly funded corporations like KSMC is whether they could develop advanced manufacturing technologies and land enough orders from clients to be profitable. It turns out that in addition to semiconductor makers, Korea needs more fabless software developers.
The proposal was introduced during a seminar hosted by the National Academy of Engineering of Korea (NAEK). The plan aims to address structural weaknesses in the industry, such as an over-reliance on Samsung's advanced nodes under 10nm amid the lack of mature process technologies. Smaller system semiconductor firms struggle to thrive as Korea lacks manufacturing diversity, as seen in Taiwan, where companies like UMC and PSMC that focus on mature and specialty nodes complement TSMC's advanced process technologies.
The Semiconductor Industry Association's Ahn Ki-hyun called for a long-term government investment. Experts project that an investment of KRW 20 trillion ($13.9 billion) in KSMC could result in economic gains of KRW 300 trillion ($208.7 billion) by 2045. However, the big question is whether $13.9 billion is enough to establish a chipmaker. Another concern about publicly funded corporations like KSMC is whether they could develop advanced manufacturing technologies and land enough orders from clients to be profitable. It turns out that in addition to semiconductor makers, Korea needs more fabless software developers.
The proposal was introduced during a seminar hosted by the National Academy of Engineering of Korea (NAEK). The plan aims to address structural weaknesses in the industry, such as an over-reliance on Samsung's advanced nodes under 10nm amid the lack of mature process technologies. Smaller system semiconductor firms struggle to thrive as Korea lacks manufacturing diversity, as seen in Taiwan, where companies like UMC and PSMC that focus on mature and specialty nodes complement TSMC's advanced process technologies.
This is a good idea... (Score:5, Insightful)
IMHO, having more chip fabs, each geographically distributed, is a very good idea. Being able to fab silicon is as important these days as steel, cement, and basic manufacturing. If a developed country can't make a modern car by itself, then they need to consider things, as one never knows when allies will leave or collapse, or supply lines get cut, either by climate change or by geopolitical shifts.
It definitely wouldn't hurt South Korea to have their own KSMC, because it brings more fab capacity to the world, and if one looks at TSMC, they are backed up for years on a lot of production lines, so there is always some market for both CPUs at sub-nanometer process nodes, as well as MCUs at 20nm process nodes which are dirt cheap and high yield.
As an added bonus, it can't hurt to have more process node technologies in the market.
Re: (Score:2)
Having more chip fabs, each geographically distributed, is a very good idea.
During Covid, we had chip shortages, and governments did what they always do and massively overreacted.
Now, we are heading into a capacity glut. Now is not a good time to build fabs.
as important these days as steel, ...
Steel mills are huge subsidy sinks. Only specialty mini-mills make money.
If a developed country can't make a modern car by itself, then they need to consider things, as one never knows when allies will leave or collapse
America's biggest source of imported cars is Mexico. The second biggest source is Canada.
If Canada tries to cut off our car supply, we can parachute the 82nd Airborne Division into Laurierville and seize their maple syrup reserve.
Re: (Score:2)
"If Canada tries to cut off our car supply, we can parachute the 82nd Airborne Division into Laurierville and seize their maple syrup reserve."
Ah, the el Bunko theory of international relations, i.e., screw our allies because America First....and soon the U.S. has no allies, and preserving the relatively stability of the World necessary for American business is left to the tender mercies of the CCP, Putin and his Kelptocrats, the Iranian MuffinHeads, and that totally stable genius running the N. Korean inte
Re: This is a good idea... (Score:2)
Applications using cutting edge chip tech are not that crucial to society. We can live happily without.
Time and money (Score:3)
The biggest problem here is that it take not only money but it takes time to a competing fabrication tech which means making a sustained and repeated investment. However, fabrication tech is a moving target which means you will be playing catch up for many years. However, process node size is just one aspect. Due to the scale, it's become more important now to decrease transistor size and (in my opinion) moving toward a 3D fabrication approach.
However, I think that chasing the cutting edge is a bit silly when the 24nm process is robust and highly useful. The world is made out of a lot more ICs than just the fastest ones which are almost entirely wasted anyway. Even x86 server chips are moving toward executing in parallel rather than executing as fast as possible. To me, this signals that there is a need for a paradigm shift in how server chips are designed which leaves open the possibility for chips made with larger processes that completely ditch the highly wasteful speculative execution and branch prediction.
Re: (Score:2)
However, I think that chasing the cutting edge is a bit silly when the 24nm process is robust and highly useful.
This all depends on the the use case. For many use cases, functionality, reliability, and cost are key, and 28nm is fine for those.
If very low power for a premium device is needed, then a leading edge process may be justified. If a system needs to be scaled out like a supercomputer, then a leading edge process is likely also justified.
Re: (Score:2)
Do both, like China is. Ramping up manufacturing for older chips and process nodes, while doing the work to develop the next generation of high end manufacturing.
Korea has the right idea. Private industry can't deliver this.
Re: (Score:2)
more important now to decrease transistor size
The problem with smaller transistors is heat.
toward a 3D fabrication approach.
The problem with 3D fabrication is heat.
24nm process is robust and highly useful.
Sure, but tonnes of older fabs can make those, so there's little profit.
TSMC is in an entirely different market.
Re: (Score:2)
The problem with smaller transistors is heat.
The problem with 3D fabrication is heat.
Of course. That doesn't mean there are no solutions, it just means more research is needed.
Sure, but tonnes of older fabs can make those, so there's little profit.
If the point is only to generate profit then they are doing it for all the wrong reasons.
Uh (Score:2)
Can't we put this stuff in a country that doesn't have a major dispute with neighbors? Taiwan could be decimated any day in a war with China. South Korea could be decimated any day in a war with North Korea. Let's not put the world's most important technology assets in an unstable region FFS.
I'm sure we could get the world's top semiconductor physicists and engineers to move from Taiwan, South Korea, and wherever else to say Switzerland or something. Or maybe Australia.
Corruption? (Score:2)