Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
?
EE Times-Asia > Optoelectronics/Displays
?
?
Optoelectronics/Displays??

Micron CEO: Of memories and foundry models

Posted: 01 Jun 2007 ?? ?Print Version ?Bookmark and Share

Keywords:Micron's diversification strategy? fab-lite strategy? DRAM market?

Appleton: "I have not heard anybody predict a 'killer app' with any accuracy in the last 10 years."

For some time, Micron Technology Inc. has grabbed its share of headlines. Micron was once a mere DRAM player. Over time, the company has expanded into CMOS image sensors and specialty memories, and formed a NAND flash venture with Intel Corp. As part of its diversification strategy, Micron also jumped into the end-user market by buying Lexar Media Inc.

What will Micron Technology do next? To get a handle on the company, EE Times caught up with Steve Appleton, its president, CEO and chairman. In the interview, Appleton addressed a multitude of topics, including the state of the industry, the foundry/fab-lite models, private-equity buyouts and if Micron is a candidate for a leveraged buyout. He also talked about the company's diversification strategy, universal memory, and whether he wanted to buy the NOR businesses of Intel and STMicroelectronics.

EE Times: Micron has been brutally honest about the state of the DRAM and NAND markets. Can you elaborate?
Steve Appleton: We tend to be pretty straightforward on how we see the market. The market is weak now. Even though DRAM might be down 10 percent to 15 percent in average selling prices (ASPs), that's not catastrophic for us. Our DRAM business is doing pretty good, but it is weakening. We have some weakening in DRAM ASP. And NAND is just under incredible pressure.

Some say there are no 'killer apps' to drive chip growth. Do you see any on the horizon? Is Vista a 'killer app'?
Vista will be helpful. But even when MP3 was starting to consume NAND, nobody called it a "killer app" until after it happened. I have not heard anybody predict a "killer app" with any accuracy in the last 10 years. I think we're in a period when the applications haven't caught up with the price elasticity of the product.

What's the geographical breakdown of Micron's sales these days?
If you break it down, about 20 percent is in Europe and 40 percent in Asia. When you think about the growth areas, we are first focused on the Far East, and that's where the action is. The U.S. market is not dead; it's still healthy.

What's the fastest-growing market?
Probably China. About 11 or 12 percent of our revenue comes from China aloneand it's growing. China has a middle class now of 300 million people, which is basically the equivalent of the entire U.S. population. The other interesting part about it is that they still haven't come up on the per capita GDP scale. They are still running at $2,500 now, but it was $2,000 last year. They still have a lot of growth ahead of them.

There's been a new wave of private-equity buyouts in the IC industry. Why?
I think there are three factors you have to consider. Being a public company has an enormous burden today. In fact, you're challenged to make long-term decisions and to sacrifice short-term profits. The second factor: There is a lot of money in the private markets. They have a tremendous number of resources at their call. The third piece is that the industry is maturing.

Is Micron interested in going the private-equity route?
There are some models that work for private equities, and some that don't. Looking at it, I don't think Micron is a good model for it because of the volatility of the markets and the capital required. We're not in any discussion stage about it.

A lot of leverage comes from optimizing the company and its model. As an example, NXP came out of Philips. The entity, which is a very small part of a much bigger company, lent itself to having inefficiencies. You can go in and clean that up. I don't know if that is going to make the deal worthwhile, because they paid a lot for it.

If you look at an example like Freescale, they were clearly a company that came out of Motorola. They had a lot of inefficiencies. But Freescale was on its way to cleaning that up before it was acquired.

Many IDMs are moving towards a fab-lite model, such as Freescale, and to some extent, TI. Any comments?
If you don't have a business that drives volume, you won't be able to afford your own fab. That's not complicated. In TI's particular case, they rely much on external foundry. I don't think it was a volume issue as much as a duplication of R&D. Although their decision alarmed a lot of the IC world, they are very much embedded with Taiwan Semiconductor Manufacturing Co. Ltd and United Microelectronics Corp. They have too much volume with their foundry partners, and so it was a pretty rational decision for them.

Any comments on the foundry model in general?
It's a model that allows an aggregation of those who couldn't do on their own. It could be a dangerous model in terms of your IP. They are servicing a lot of different customers. They are trying to take their accumulative knowledge and make it better for everyone by attracting more customers. It's a fine line.

Not long ago, Micron began its diversification strategy, jumping into CMOS image sensors, specialty memories and NAND. Is the company changing course given the soft environment?
We're not changing course at allwe're driving forward. We are very optimistic about the market and about the applications that will consume our products over the longer period of time. We're completely committed to the path we're on. We always knew that we would hit speed bumps in the road. We always knew that NAND would behave like DRAM at times and that the mobile market would have to flatten in penetration.

Micron is moving toward an integrated IDM model. The company has IC design, a photomask shop, fabs and assembly plants. Micron bought Lexar, a supplier of flash cards and USB drives. What's the rationale behind that model?
We've done a lot of things on the partnership side. We've had deals with Intel and Photronics, and we joined IMEC. We recognize that we can't do everything ourselves.

Micron formed a joint photomask venture with Photronics. Why even have a mask shop in the first place?
The reason you do it is because mask shops have become generic. They don't have the capability to develop advanced masks to the specs we need. So we had to do it ourselves. Our mask costs, by the way, went down as we went with our own mask shop and our turn times had something like a tenfold improvement.

What's the status of Lexar? Why did you buy them?
It's going great. The integration has gone well. We did restructuring and integrating. That was a little uncharacteristic for us, because we had historically said that we would not compete with our customers. And as soon as we started going into the marketplace in the channels with a branded product, we started doing that. Now, we do it very cautiously, because we don't want to compete with them.

Is there a missing piece in Micron's strategy?
Yes, there is a missing pieceit's packaging. We do a lot of packaging. We have a big assembly operation. But in an integrated model, we are missing parts of the packaging for the next generation of applications. In particular, when you think how that translates to the imaging business, you think about wafer-level modules, lenses and camera module integration. We have a lot of partners for that.

So do you acquire a company or bring more expertise in-house?
We will do both. We will acquire companies that have expertise in the area. At the same time, we will also have to continue partnerships.

There is a debate in terms of how long the industry can extend the floating gate in flash memory. Any comments on that and 'universal memory'?
We think the floating gate is fine down to 35nm. Then, we talk about different materials, which are really just state changes, such as phase-change, Ovonyx stuff or related to it. This is still several years away. We've done work in all of that stuff. We have tons of MRAM technology IP. We did lots of work in Ovonyx. We've done lots of work in phase-change RAM, which is basically phase-change materials.

If you back up five years ago, they said it was the holy grail. It was supposed to be a non-volatile memory that would replace everything out there. Of course, it didn't materialize, and it's been years later now. In fact, someone tried to convince me that Ovonyx was going to be introduced in the market in 2000. In 1999, they said: "You got to do something." It's going to take over the world. Here we are years later, and what's happened?

The discussions are more directed now at what parts of the memory market it is aimed at. Is it really a better solution at embedded? Or does it attack the NOR market first as opposed to the NAND? You don't really even hear about it replacing DRAM anymore.

- Mark LaPedus
EE Times




Article Comments - Micron CEO: Of memories and foundry ...
Comments:??
*? You can enter [0] more charecters.
*Verify code:
?
?
Webinars

Seminars

Visit Asia Webinars to learn about the latest in technology and get practical design tips.

?
?
Back to Top