Sign In

A story about semiconductors: Samsung Electronics and SK Hynix

Haebom
Recently, over a meal with someone from Samsung Electronics DS, we started talking about a company called Padu, which has become a hot topic lately. The conversation turned to semiconductor design and production; although my knowledge is limited and I couldn't understand everything, I learned that among Padu and the other new domestic semiconductor manufacturers being discussed, only one actually has its own proper production line. The rest just do the design work and then have small batches produced as OEMs on other production lines.
https://www.sammobile.com/news/nvidia-beats-intel-samsung-tsmc-biggest-chip-brand-q3-2023/
These days, in the AI chip industry where prices are sky-high and everyone is lining up for a ticket, Nvidia and TSMC hold incredibly strong positions, having already built substantial barriers to entry for any would-be challengers. Nvidia controls about 95% of the GPU market, and TSMC holds roughly 90% of the world’s advanced chip sector. Both companies boast high profitability, with operating margins of 74% and 60%, respectively.
The powerful position of these companies comes from several factors. For instance, Nvidia owns thousands of patents covering hardware and related software libraries, while TSMC has secured its edge over competitors like Samsung through more than eight years of investment in advanced packaging technologies. On top of that, the extreme ultraviolet lithography machinesessential for making advanced chips are exclusively supplied by ASML in the Netherlands, and even if you place an order now, it typically takes about two years on average to receive one.
미국에 상장된 반도체 관련 주식들 가치입니다. 해당 수치와 삼성전자 반도체 부분, sk하이닉스 시총이랑 비교해보면...
Given all this, for new chip manufacturers to enter the market requires huge capital and time, and even if they clear technical hurdles, catching up with the advantages that Nvidia and TSMC hold on the ground is a major challenge. So far, domestic companies have positioned themselves more around general-purpose, frequently used chips rather than high-performance ones. Of course, demand for these isn't about to drop off right away. Still, unless a strong consortium is formed or a new breakthrough found within the next two to three years, it looks like it will get tougher to find success in semiconductors going forward.
Subscribe to 'haebom'
📚 Welcome to Haebom's archives.
---
I post articles related to IT 💻, economy 💰, and humanities 🎭.
If you are curious about my thoughts, perspectives or interests, please subscribe.
haebom@kakao.com
Subscribe
2
Haebom
Haebom
엔비디아가 3분기 마이크로소프트와 메타에 각각 15만대의 H100 AI GPU를 판매했다고 한다. 이에 반해 구글 아마존 오라클 텐센트는 각각 5만대. 코어웨이브 4만대, 바이두 3만대, 알리바바 2.5만대.
네이버가 얼마전 연 세종시 '각' 데이터센터는 엔비디아 A100 텐서 코어 GPU 2240개로 구성되었다. 이걸 4로 나누면 서버 560대라는 소리다. (하나의 서버에 칩 4개 꽂는다. SK텔레콤이 연초에 AI 인프라 두배로 늘렸다고 보도자료 뿌렸는데 알아보니 65대 있던 거 65대 늘린 거였다. 그래봐야 130대다.)
See latest comments