Opinion: Samsung misses foundry boat
(12/08/2009 2:07 $� EST)
Recently, I was in Korea for an editorial trip. One of my goals was to find out how Samsung Electronics Co. Ltd. is doing in the foundry business.
My opinion after the trip: Samsung is still learning the foundry ropes. But in reality, the South Korean memory giant has been a major bust in the foundry business and it still doesn’t really understand the arena.
For years, Samsung has been in the foundry business and claims it wants to become a major player in the arena. Samsung entered the foundry business in 2006 with sales of about $75 million. In 2007, it had sales of $385 million, according to IC Insights, although this was behind other leading IDMs that are in the foundry arena.
In recent times, Samsung claims to have won some foundry customers, including Qualcomm, Xilinx and perhaps others. Samsung is ramping its foundry business within its S1 300-mm fab, located in Giheung, South Korea. Samsung is also part of IBM Corp.’s ”fab club,” which is developing R&D technology. IBM, Chartered and GlobalFoundries are in the same group.
Still, it is difficult to find out what or how Samsung is doing in the foundry sector right now. The company has been tight-lipped about its efforts. I tried to arrange interviews with Samsung executives in Korea regarding its foundry and memory efforts, but all requests were denied. At a recent trade show in Korea, Samsung has some representatives at the event, but none from the foundry group.
In Korea, word on the street is that Samsung is finding the foundry business somewhat harder than originally anticipated. There was also word that Samsung is looking to become more aggressive.
Now, according to a recent report from a newspaper in Korea, Samsung plans to double its production of chips for foundry production every year until it rivals market leader Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC).
The paper, Chosun Ilbo, quoted a Samsung spokesman, saying the company has decided to expand its foundry production as a “future growth engine with the long-term goal of becoming as big as the world’s number one TSMC.”
If that report is indeed true, this confirms my belief or opinion that Samsung doesn’t really understand the business. Years ago, foundry companies aggressively expanded their fab capacities for the sake of bragging rights and having more fabs than others.
That strategy worked when foundries were growing by alarming rates and were immune to the downturns. Those days are over, however.
The foundry business is not immune to today’s downturns. And vendors have learned to expand their capacities on an incremental basis to maintain a healthier supply/demand balance, as compared to the ”fab- a-day announcements” of yesterday.
Samsung doesn’t get it. The foundry sector is a service business. It has nothing to do with churning out cheap and simple commodities like memories.
True. Samsung could be finding new customers and needs to expand. I suspect it will gain some share. But expansion for expansions sake is an old and desperate strategy, leaving me to believe Samsung will ultimately fail once again to find a new and growing non-memory market.