You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Samsung has had yield problems with 10nm down to 5nm. Hard to believe that 3nm GAA is going to be yielding anytime soon. It will also be interesting to see what the specs are and how close Scott's projections are:
I was asked to give a talk at the 2021 ISS conference and the following is a write up of the talk. The title of the talk is “Logic Leadership in the PPAC era”. The talk is broken up into three main sections: Background information explaining PPAC and Standard Cells. A node-by-node comparisons of...
Samsung Foundry 3nm Gate All Around Process Node, 3GAE, Delayed to 2024! This would arrive in a similar time frame to TSMC's denser 2nm GAA technology.Samsung and Intel continue to slip further behind TSMC in leading edge technology. Will they ever be able to catch up?
Samsung Foundry 3nm Gate All Around Process Node, 3GAE, Delayed to 2024! This would arrive in a similar time frame to TSMC's denser 2nm GAA technology.Samsung and Intel continue to slip further behind TSMC in leading edge technology. Will they ever be able to catch up?
Thank you for finding that article! If true, this is really bad news for Samsung Foundry! Both Daniel and Scotten already alluded to problems for Samsung in their comments so this just adds further support of their views.
If Samsung truly nerfed their 3nm GAE to the extent written, then TSMC is going to be the only game in town for high performance computing...
We will give Samsung the benefit of the doubt and assume this comparison is with their most advanced version of the 7nm class of nodes (5nm LPP). If so, then this is not as much as a stealth nerf as it is a new comparison to the latest and most advanced Samsung process technologies. Regardless, these gains may not even bring Samsung to power and performance parity with TSMC’s 6nm node, let alone the 5nm node being used in current iPhones, iPads, and Macs.
It was always in the repertoire of Samsung to race to claim node parity with TSMC just to get clients for latest node service, no matter how bad yields were.
Some times I think Samsung did manufacture latest node at loss just to appear that they are still in the race.
South Korea did recently announce that it would commit $450 billion to their semiconductor industry. Most of it is aimed directly for the advancement of their logic chip industry. It'll be interesting to see how their ambitious investment plan pans out. From a Korean point-of-view, the world is against them...and that isn't far from the truth. TSMC and Japan recently announcing an $337 million R&D project in Tsukuba, Japan. Intel's IDM 2.0 strategy (and US CHIPS Act) is seen as a threat to Samsung. China is buying semi equipment, building fabs and looks to surpass South Korea in the not-too-distant future. Global Foundries is investing $4 billion in a Singapore chip plant. The EU has stated its intent to have manufacturing capabilities below 5nm and production of 20% of the global semiconductor production in Europe.
It's hard to believe that we won't have a wafer glut in the not too distant future given the CAPEX spending that has been going on of late. We have a serious who has a bigger CAPEX contest going on.
South Korea did recently announce that it would commit $450 billion to their semiconductor industry. Most of it is aimed directly for the advancement of their logic chip industry. It'll be interesting to see how their ambitious investment plan pans out. From a Korean point-of-view, the world is against them...and that isn't far from the truth. TSMC and Japan recently announcing an $337 million R&D project in Tsukuba, Japan. Intel's IDM 2.0 strategy (and US CHIPS Act) is seen as a threat to Samsung. China is buying semi equipment, building fabs and looks to surpass South Korea in the not-too-distant future. Global Foundries is investing $4 billion in a Singapore chip plant. The EU has stated its intent to have manufacturing capabilities below 5nm and production of 20% of the global semiconductor production in Europe.
It's hard to believe that we won't have a wafer glut in the not too distant future given the CAPEX spending that has been going on of late. We have a serious who has a bigger CAPEX contest going on.
The lion share of fab capex went to advanced nodes by big foundries. The biggest capacity crunch in the last few years was on mature nodes, and you don't get new real physical capacity there since new 200mm equipment is simply not manufactured anymore. Mature node fabs will just try to squeeze a bit more from their equipment with automation, and optimising operations. Maybe they will throw in few new deposition, or etching tools, but that's all.
semiwiki.com
semianalysis.substack.com
