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!
I don't track foundry news (others know far more about this area than I do), but this news seems interesting:
YONGIN, SOUTH KOREA (Reuters) - Samsung Electronics (005930.KS) plans to triple the market share of its contract chip manufacturing business within the next five years by aggressively adding clients, a senior company executive said, as it targets new growth drivers for the chips business.
But the firm lags well behind Taiwan's TSMC in contract manufacturing: TSMC held a market share of 50.6 percent last year compared with Samsung's 7.9 percent, according to research firm IHS. It also trailed U.S.-based Global Foundries, which had a 9.6 percent share, and Taiwan-based UMC's 2303.TW 8.1 percent.
Samsung does not release foundry numbers so how do they know?
The estimated 5.3 trillion won ($4.76 billion) business at Samsung was split off as a separate arm within its semiconductor division in May, in a clear statement that the technology giant was preparing to focus on the business and narrow the big market share gap with leader TSMC (2330.TW).
Why is this estimated? Why can't Samsung be transparent here? If Samsung wants play with the pure-play foundries and claim leadership they should release real numbers like everyone else, my opinion.
Now that Sasmung LSI (internal Samsung chip development) is a Samsung Foundry customer is that counted in their revenue? Because Intel can do the same and they would be the #1 foundry!
I still say the IDM foundries that compete with their customers have limited appeal, especially for systems companies. How would you feel if you competed with Mobileye and did your chips at Intel? Achronix was one of Intel's first foundry customers then they bought Altera... Same thing goes for QCOM and Samsung, Apple and Samsung, etc...
In theory, spinning off the foundry business is meant to address the above. In practice, we know that any separation between Samsung's design and foundry business will be superficial, and Samsung has a long history of IP theft which they have been extremely blatant about.
While I'm sure it's possible for Samsung....but I wouldn't bet against TSMC.
First, they really seem to be doing quite well right now. I'm currently vacationing in Taiwan. I was at Taichung a couple days ago and I must say that TSMC has greatly expanded their presence there. I saw 2 *MASSIVE* new buildings that didn't exist last year. The taxi driver stated that the locals were shocked at the incredible pace at which TSMC built the facilities there.
Lastly, Digitimes did an interview with the GlobalFoundries CEO (Sanjay Jha)...it may well be that 5nm and below will be a two horse race (Samsung & TSMC). He didn't sound sure that GF would "get enough of a return" on 5 nm.
"Samsung Electronics (005930.KS) plans to triple the market share of its contract chip"
So they PLAN to do this, huh? And that's all it takes? You have to wonder why they didn't *plan* this ten years ago...
There used to be a game in music (before we all got wise to it) that by simply asserting that (unknown musician from faraway country) was a star, by living the lifestyle, and by occasionally making outrageous statements, you could sometimes get lucky and turn the fantasy into reality.
This appears to be Samsung's game plan here: drive to the fab in a Rolls Royce, maybe hire Frank Gehry to create a facade for the new fab, maybe place a few strategic ads on the Facebook feeds of "people who are interested in computers", maybe start a rumor that you add cocaine to your finFETs to make them switch faster, and, voila, you're a fab rock-and-roll star with everyone talking about you and eager to dance to your next process.
While I'm sure it's possible for Samsung....but I wouldn't bet against TSMC.
First, they really seem to be doing quite well right now. I'm currently vacationing in Taiwan. I was at Taichung a couple days ago and I must say that TSMC has greatly expanded their presence there. I saw 2 *MASSIVE* new buildings that didn't exist last year. The taxi driver stated that the locals were shocked at the incredible pace at which TSMC built the facilities there.
This is certainly good news. It confirms what we are seeing with AI readership, it is touching all applications. Notice that the systems companies are chiefly behind this (Apple and Google). Now that TSMC is closely collaborating with customers on packaging this will make their processes much more sticky, yes?
I would not assume Samsung executives are brash or naïve, regarding this statement. And I would not assume that the lever to produce this result has been revealed yet.
It's not a bad assumption that Samsung is targeting TSMC, but, the article correctly points out, before that happens, first GF and UMC have to be dealt with. So I think the real targets are GF and UMC.
I would not assume Samsung executives are brash or naïve, regarding this statement. And I would not assume that the lever to produce this result has been revealed yet.
It's not a bad assumption that Samsung is targeting TSMC, but, the article correctly points out, before that happens, first GF and UMC have to be dealt with. So I think the real targets are GF and UMC.
I agree. UMC is now shipping 14nm and SMIC should not be far behind. I do not see them going after 7nm aggressively though. GF is going after 7nm and has already started 5m so GF is the one Samsung will need to beat to get the "Not TSMC Business". I really do not see Samsung beating TSMC head-to-head anytime soon. The TSMC yield is so much better and they have a HUGE advantage over Samsung on the business side.
BTW, this year is TSMC's 30th anniversary. There is quite a gala planned in Taipei in October....
