By Bruce Einhorn In its race against bigger DRAM companies Samsung Electronics of South Korea and Micron Technology (MU)
of the U.S., German chipmaker Infineon Technologies is counting on Asian alliances. In late February, Infineon (IFX) President and CEO Ulrich Schumacher traveled to Asia to meet with executives from Nanya Technologies, the Taiwanese semiconductor company that's Infineon's new joint-venture partner. At the same time, Schumacher met with Taiwanese officials to sort out problems with Mosel Vitelic, Infineon's old partner in a joint venture called ProMOS Technologies (see BW, 3/24/03, "Why Infineon Wants Asian Allies").
I recently met with Schumacher and asked him about ProMOS, Nanya, and another partner in China. Edited excepts of the conversation follow:
Q: What is the ProMOS endgame?
A: "Endgame" is such a nasty word. Of course it's a pity that we break up, but it happens in business. Mosel Vitelic changes its vision of the business -- they're no longer a long-term committed DRAM [dynamic random-access memory chips] partner. We had to decide on the next level [investment], and we want to do that with a partner with the same commitment. It's clear our strategies don't fit anymore.
Q: What went wrong?
A: ProMOS was one of our fabs [or chip-fabrication plants]. It was part of Infineon's fab management. A fab should drive productivity, making [production] cheaper and more efficient. That's what a fab does. But for Mosel, this was also a financial investment. We never cared about ProMOS's market cap. It's a fab. We don't care. But they started to say that ProMOS had its own image, and they were tempted to make ProMOS more than a fab.
Q: What's bad about that?
A: The [risk is that] complexity goes up, and the productivity goes down. It [can] defocus [the fab's] activity. They wanted to try to make ProMOS look like an independent semiconductor company so they could get the highest value. We just want to have an efficient DRAM fab.
Q: So what happens to your stake in ProMOS?
A: We are balancing between trying to safeguard the future of ProMOS, maximizing ProMOS, and defending Infineon's interest.
Q: Would you be willing to sell to a rival like Elpida? Or even to Samsung Electronics?
A: I would be the happiest man on earth if Elpida wants to walk in. That would get my support. I'm not afraid -- whoever wants to come in, I have no problem with that. I just want my money back -- or my output back.
Q: Given the problems with ProMOS, why should we believe that your new joint venture with a different Taiwanese partner, Nanya Technology, will go any better? I'm reminded of the old joke about second marriages -- that they're triumphs of hope over experience.
A: Strong characters do marry a second time. It's a question of age and maturity. If you marry a second time, you're more selective in your choice. You've learned a bit about how to make things better.
Q: And how does Nanya differ from Mosel?
A: This partner is completely different. Mosel in 1996 had no DRAM technology, no competence in this field. Nanya today is a proven, very capable DRAM company. It came from nowhere to be No. 5 in very few years. You can see their efficiency in the fab and how they're behaving in the market. Nanya belongs to the Formosa Plastics group, so it's financially pretty solid. So we've had a much better start.
Q: But isn't there a potential conflict here, since separate from your joint venture both Infineon and Nanya are competing for the same DRAM customers?
A: In the DRAM market, our rival is Samsung. Nanya knows, as we know, that by no means does it make sense to fight each other. We have to develop technology or we can't survive, so we need a partner.
Q: Why Nanya, then, and not Micron?
A: Micron has a different business model. It's a fast follower, buying fabs all over the world. Whenever the spot market is high, it shifts [sales] to the spot market and kills the OEMs [original equipment manufacturers]. We always give the OEMs what they want -- even if the spot price is higher.
Q: Any other big Samsung rival?
A: The rest are [really struggling]. Hynix, Elpida, you name it. They didn't make sense.
Q: You've started working with Chinese foundry Semiconductor Manufacturing International in Shanghai. Why?
A: China is this big phenomenon. Is it the biggest market of the future? Or is it the biggest threat? Nobody has a clue what China really is. What do you do now? You can say, "As long as it's vague, I'm not going to participate." The risk is that you lose a huge market and [pool of] engineers.
For cowards, that would be the right strategy. But if you want to be a leading semiconductor company in the world, there's no way you can ignore China.
Q: Yet many people say China does a terrible job protecting intellectual-property rights, so it's better to stick with low-end, older technology in China to avoid getting ripped off.
A: That's the half-pregnant approach. You have sex, but don't have sex. You need a special kind of thinking to really believe that. If you go to China, you have to go full way. There's no way to say "yes, but...." We have to go aggressively. If you don't do it, someone else will do it.
Q: What happens then if someone in China takes Infineon's technology and runs off with it?
A: The good news of the DRAM business is that innovation is fast, you can't run far. One year from now, the next generation will be 30% more productive, and two years from now 50%. So in two years time, you're out.
Q: So there's no reason to worry?
A: You can do things on paper to protect your technology. But ultimately, you can't. You have to build in the possibility that things will happen. I'm not so afraid.
Q: Why not?
A: This is a business where you have to be a little optimistic. The DRAM business is not a business where you can hide behind a curtain. Einhorn covers technology from Hong Kong for BusinessWeek. Follow his weekly Online Asia column, only on BusinessWeek Online