Two years ago, Agere Systems Inc. began transforming itself into a fabless semiconductor company as part of a wide range of reorganization measures it took to restore profitability. Peter Kelly, Agere executive vice president, global operations, recently discussed the changes since implemented by the company and the outlook for the semiconductor industry with Electronics Supply & Manufacturing executive editor Bolaji Ojo.
ESM: What's the latest on Agere's fablite plan:
Kelly: 2004 and 2005 would be very much transition years. Most of our sales this year would be supported from our two internal fabs. Probably 70 percent would be from internal fabs and 30 percent from external sources. And as we go into next year, it drops to 50 percent internal and 50 percent or even 60 percent external. It's really 2006 before we become, for all intent and purpose, fabless. We would like to become 100 percent fabless but one simple way to think about it is that any design that we have that is less than 0.14"0.13 and beyond"is completely external.
We have no internal capacity in the 0.13 micron and below range. Most of our new designs are actually at 0.13 micron so by the natural lifecycle of the products, future growth would increasingly come from foundry supply. At 0.14 micron and above, we'll source from our fab in Orlando, Florida, and our joint venture fab with SMP. We are not making any significant investment in capacity in those fabs. As demand has exceeded the size of those fabs, we've gone out and bought those products mainly from Chartered Semiconductor and from SMIC in China.
ESM: What is your relationship with TSMC?
Kelly: We have a very strong relationship with TSMC. Our main 0.13-micron foundry partner is TSMC. Chartered is also our 0.13 foundry partner. For future technology needs, that is, for 0.13 micron and beyond, it's TSMC and Chartered. For our current CMOS technology, 0.14-micron and 0.30 micron, it's Chartered and SMIC. We started using SMIC in the last 12 months and we have a hope that the company would become a successful foundry and that we'll be able to use them in future. You can divide our demand into three areas: new technology, which is 0.13 micron; legacy technologies, which is 0.14 micron to 0.30 micron and; then boutique processes, for instance, some of the analog processes.
The boutique processes we do with companies like Jazz Semiconductor, which was spun out of Conexant. They tend to be small volume but special processes. The new stuff is in Chartered and TSMC but although we don't see any substantial growth from the legacy technologies we do think they will be around for quite a while.
ESM: Do you have a timetable for when Agere might take its last two fabs offline?
Kelly: No. They are both fully automated 8-inch fabs. They are nice fabs and from a capacity perspective they produce about 25,000 wafers a month. That's one of the reasons we didn't add any extra capital. We can use that capacity for quite a while.
ESM: Agere's CEO, John Dickson said recently that the company had successfully negotiated lower pricing with its foundry partners. Was this due to consolidating wafer procurement at certain foundries?
Kelly: I think you have to take it in the context of the fact that in our industry you have to drive down your cost every year. Our customers expect us to reduce our prices for them every year and we do the same with our suppliers. It's not in the context of the DRAM world where there is a confrontational set of price agreements"when times are tight the DRAM guys put prices up and when times are not so tight they are forced to reduce prices.
The relationship we have both with our customers and with our suppliers require that over extended periods we work together to drive down cost. And I think what John was responding to was a question on the availability of capacity. What John was trying to get across was that we are not in a spot market fight with our suppliers.
We are in long-term relationships where we work together to provide the best possible deals to the customers and for us all to make money. There is no point in putting a supplier out of business and there is no point in a customer putting us out of business. Because of the types of products we make we have extended multi-year relationships and we've learned that there is no big advantage in trying to take some sort of short-term advantage.
ESM: Is Agere signing long-term, higher volume wafer-sourcing agreements with some of your foundry partners?
Kelly: Absolutely. With Chartered and TSMC in particular we look several years out. It's typical for us to agree quantities but we do also agree on shares. With both of those suppliers, we've given them contracts where they are going to be the source of certain products. Because it's very difficult for us to move a product from one foundry to another once we've contracted a product, the foundry is pretty sure that, as long as the customer is successful, it's going to get very substantial revenue.
ESM: The fablite strategy is a relatively new one for Agere although you've been doing this for several quarters now. What have you had to do differently to support the strategy?
Kelly: It is a new strategy for the company but we have done some work with the foundries for many years. We have relationships with Chartered and TSMC that go back five to seven years. But having said that it is a big change for us because it requires a degree of optimism with the supplier that I don't think we are used to. It requires a massive amount of trust because you have to share lots of information with each other to be successful.
It's easy to say we want a foundry supplier to be a strategic supplier but what we are trying to do is position ourselves as a strategic customer to them. We know we are going to buy the products from them but we also want to position ourselves as a strategic customer to them. We want the foundry to feel that we are one of their top customers. There are different things we have to do to enable that kind of thinking. Volume helps but it's not just about volume.
ESM: How does a fabless company ensure it gets its required allocation from the supplier?
Kelly: Smaller companies have a problem. At a big foundry, companies that account for 20 percent of volume take up 80 percent of the production control things like engineering time, scheduling and all the things that drive you nuts. Those very small guys are the ones that are most likely to get squeezed, as times get tough.
I think there's always the possibility that even Agere could get squeezed when there's under capacity so it would be wrong to say never. But part of the solution is having the right relationship with the foundry partner. We spend a lot of time talking about what our demand forecast is, I meet with their executives once a quarter, I have people on site in their factories working with them on a daily basis.
If you are a large company, there are a lot of things that you can do to help your suppliers and make life easier for them. As a result, you are more likely to anticipate some of these big changes. Part of a successful fabless strategy is that you have to work with the most successful companies in the world.
ESM: How do you ensure that kind of relationship if you are a small fabless semiconductor company?
Kelly: The key is technology. You have to be able to convince the foundry that you have a piece of technology that is going to potentially grow very rapidly. But it is kind of difficult.
ESM: Is there pressure on suppliers to pass cost savings on to customers?
Kelly: There are customers who say to us 'if we are to be successful, we need to cost reduce our products. How can you deliver a cost-effective solution for us?' So we'll go in and we'll look at things like can we combine two chips into one, can we make the chips smaller, and can we go into a more cost-effective package. We'll do that with our customers and then we'll go and do the same with our suppliers or our suppliers will do the same with us. They'll say 'can you use this process instead of that process, can you reduce the number of processes you have as a whole, can you reduce the number of packages you have as a whole, can you aggregate volume so we can reduce our costs.' So it's a very collaborative set of discussions. It's not combative at all.
ESM: What's the pricing situation like now?
Kelly: It's as tough as ever. You would expect now that we are in an upturn that all the components guys can put their prices up but that is not true. We have just as much price pressure as we ever did. Our customers are trying to expand their markets and the best way to expand your market is to lower your price so that in the end consumers would start buying more of your products. We are on a treadmill of constantly reducing cost.
ESM: Agere has always focused on a core group of 30 customers. Are you now going outside of this group?
Kelly: Not at all. We have about 30 customers that generate about 65% of our revenue. We still think that the approach that would be successful is to have the best products in the world and deliver it with a customer-service model that is better than anyone else. Let's say that we have four customers that generate $1 billion worth of sales, we don't want to move to 8 customers generating $1 billion worth of sales. We'll like to move to eight customers generating $2 billion of sales.
We are not a standard-products company and in order for us to give the level of service that we want, it really only makes sense for us to work with people that we think would be ultimately very large. They may not be very large today but if we believe they have the potential to be very large we do spend a lot of time with them. From a sales perspective, they'll have their own account team that will spend time with them to make sure they understand our products and how we can help them. They'll have dedicated engineering support. We'll have custom supply chain support for them and they'll have access to our executive team.
ESM: What's your assessment of the current market situation?
I think we are in the first part of an uptick. Things are beginning to improve across the industry now. I do think you'll see blips from one month to the next. When things fall off they don't fall off in a complete linear fashion. The industry will see some good quarters and some poor quarters and then it will pick up again. I really think we are at the start of a period of growth.
I do think that Agere has moved from a surviving mode to a thriving mode. We've had four quarters of growth: 18 percent year-on-year. We've restructured the company such that even if there's a big downturn now we won't be as impacted as we were in 2001 and 2002. It's kind of nice to be in the semiconductor industry at the moment. That's something you can only say every few years.