Silterra Malaysia

URL: PDF version

Despite the attractions of the fabless-plus-foundry model, any foundry hoping to compete in an industry dominated by the likes of UMC and TSMC will need to think carefully about its strategy. In the case of Silterra, whose production facility is located in the Kulim High-Tech Park, Malaysia, the strategy combines carefully timed moves to deep submicron with the cultivation of mass-market customers. The result is financial stability after troubled times in the semiconductor industry, and strategic relationships that include a partnership with Taiwan design-service house Goyatek. Silterra is also involved in a joint development project with IMEC in Europe. According to Steve Della Rocchetta, EVP, Sales and Marketing, Silterra is going into 2006 with optimism.

DigiTimes.com spoke recently with Della Rocchetta about Silterra's strategy, technology and prospects. This interview also includes contributions from Koh Meng Kong (MK), director of Marketing for Silterra in Asia.

Q: Do you have any comments on the difficulty of designing and implementing RF and mixed-signal devices in CMOS process? Are there still special challenges in these areas, or is this kind of design and fabrication now pretty much standard?

A: When you design for mixed-signal and RF-CMOS devices, the quality of the model you use is very important. Here I'm referring to the SPICE models that designers work with. We believe that one of the factors that differentiates Silterra is the quality of our RF models. We believe that it's among the best, if not the best model available for RF technology, particularly when you're look at it from the point of view of signal-to-noise ratios as
you extend the gigahertz range. We try to listen to what designers say and then give them better models with which to work, so that their silicon has a better chance of having first-time success.

MK: Our RF model has in fact been fully characterized to 10GHz, and we have demonstrated the model at 10GHz. That's good enough for wireless LAN, for example, and it goes beyond what most players in the RF market can supply.

A: Most players in the RF market only characterize their designs up to 5GHz.

Q: I'm getting a very strong message from you that your strategy is to focus on mass markets. Clearly RF is becoming a mass market. What about high voltage and mixed signal?

A: High voltage is a huge market. There are two distinct market segments. You have large-panel displays, where the application would typically be digital TV. Then you also have small-panel displays, and these are typically used for cell phones – as either the main panel or the sub-panel.

I believe that the small-panel market for cell phones represents more than a billion displays per year, when you total both main and sub panels. We are very focused in that area, and we have industry-leading technology. We have 0.22-, 0.2- and 0.18-micron technologies in production and more advanced 0.16- and 0.13-micron technologies on the horizon. We're defending this market very aggressively.

Q: Can you comment on your competitive position in those markets?

A: I think we're very well positioned competitively. We work very closely with customers to jointly define process specifications in order to provide them with the most cost-effective silicon solution. As an example, we fine-tuned the SRAM cell to 3.99-square-micron for our 0.18-micron technology.

Q: Would you like to make any further comment on your mixed-signal technology?

A: Virtually all designs today incorporate analog to some degree, so we don't really differentiate between CMOS logic and CMOS mixed signal. To us those are very much the same, and of course, these are gigantic markets. We're there at 0.18 micron, moving into 0.13 micron to maintain a competitive edge, and it's strategically important for us to do that.

Today, CMOS logic is a fast-driving market. If you look at companies such as Qualcomm, Broadcom and Marvell, these are in many ways the companies that are driving the technology. They're adopting 0.13-micron, and 90- and 65-nanometer processes in rapid fashion. If you don't do 0.13 micron and 90 nanometers, you're going to be relegated to the backwaters of that market.

Q: What would be the backwaters, today?

A: Probably 0.25 micron today, 0.25 micron and above for that market.

If you look at the predominant designs today, not in terms of the square inches of silicon they consume, but in terms of the actual number of designs, 0.18 micron is still the number one designed-in technology, and by quite a wide margin, with 0.13 micron starting to move up. So as far as CMOS logic and mixed-signal are concerned, there are very few designs being done today at 0.25-micron and above.

Q: In pursuing mass markets, would it be true to say that your broad strategy is to try and collect, at lower pricing, customers who would otherwise be picked up by TSMC and UMC?

A: It's fair to say that's part of our strategy, yes. Particularly when you are in the multi-sourced customer-owned tooling (COT) market, there's almost no getting around that. Obviously, we intend to move to a primary-source strategy, but that takes time. I have a saying, "No foundry manager ever got fired for selecting TSMC." That's a kind of variation on the old saying, "No IT director ever got fired for choosing IBM!" So for a newcomer like us, it really takes a while to achieve a level of trust where you can begin to be the initial source or the primary source of design work. It takes a while.

Q: That would also be a question of yield and right-first-time silicon, and so on?

A: Exactly. A smooth factory operation, a good cycle time and good fab execution, these are obviously crucial. You have to do it right every time, and you have to be able to do it in under four weeks.

Q: Can you sketch in Silterra's financial state of health, profitability, business strategy and so on?

A: In 2004, we were EBITDA positive, and likewise for 2005 and 2006. We're optimistic about 2006. It's always hazardous to predict the future, but it seems to me that about every five years a silicon shortage comes along, and it appears we are due for another shortage in 2006. If you look back at history, the last time that happened was in 1999-2000, and previous to that was 1994-1995. I believe that if the foundries can control their rates of
capital investment, and business continues on its current trajectory, next year could be a very good year for us and for the industry in general.

In the meantime, I can say that we control our costs very well. We don't add capital until we need it, so we have a good cost structure.

Q: Your production base is in Malaysia. How has that worked out for Silterra?

A: I think it's a good strategic location, simply because of the predominance of back-end operations in Malaysia and the Penang area. Apart from that, we also have the same kind of incentives that you would get in Singapore or in Taiwan, so I would say that the playing field is very flat in that regard.

Labor isn't really a big issue for wafer fabs. Our labor costs are in line with those of everyone else. It really comes down to your capital costs and your operational efficiency. These are the things that really separate the men from the boys.

About Silterra Malaysia Sdn. Bhd.:

Market demand driven, SilTerra Malaysia Sdn Bhd is a semiconductor wafer foundry offering major foundry compatible CMOS logic, high-voltage and mixed-signal/RF technologies down to 0.13-micron feature size. This includes complete, competitive contract manufacturing for fabless and IDM customers’ designs. SilTerra’s wafer fab has a design capacity of 40,000 eight-inch wafers per month.

Environmentally vigilant, SilTerra delivers award winning, world-class performance to its customers seeking flexible capacity, competitive advantages and around the clock customer support. SilTerra is ISO 9001:2000 and ISO 14001 certified. Founded in 1995, the company’s headquarters and factory are located in Malaysia’s Kulim High-Tech Park, and SilTerra has sales and marketing offices in San Jose (California) and Hsinchu (Taiwan). For additional information on SilTerra or its services, please visit www.silterra.com.

Silterra Contact

Koh Meng Kong

Tel: +6-012-491-0425 (Cell phone)

Tel: +604-401-4166 (o)
Email: mengkong_koh@silterra.com

Lu Ping Chiang
Tel : +886-3-574-1587 (o)

Email: pingchiang_lu@silterra.com