MUNICH, Germany Once a typical integrated device maker (IDM), Infineon Technologies AG is transforming itself into a fab-lite manufacturer Ė using foundry partners for a large proportion of its digital manufacturing and uses its own legacy fabs for RF, mixed-signal and more esoteric manufacturing. Christoph Hammerschmidt discussed the ramifications of this strategy with board member Hermann Eul.
EE Times: In the chip business, besides semiconductor expertise, application expertise is becoming increasingly a decisive competitive factor. How does this affect the value chain and what Infineon does?
Hermann Eul:: This process not only affects the value chain but is driven by the value chain. Today, we integrate functions and function blocks on a chip that, a few years ago, would have been allotted several chips. As a consequence, the interaction between these blocks is no longer in the realm of the appliance designer but belongs to the chip designer. This means that such a chip can only be designed competitively if the designer has expertise in different areas such as RF, baseband and application functionality. Someone who does not have the application expertise ultimately cannot build a modern chip - simply for the fact that he does not know what kind of functions to integrate.
EE Times: Which means that IDMs need to be responsible for process technology, chasing ever-smaller geometries. What about Infineonís decision to refrain from in-house manufacturing beyond a certain geometry node?
Eul: In part you are right. The semiconductor industry as a whole wonít get around doing the research and investing in advanced technologies. On the other hand, as a single company you nevertheless can decide in which areas to keep pace with technology progress. Anyway, one needs always to have access to the latest technologies. It is important for a company to arrange that access in a way that can turn it into a competitive advantage.
EE Times: Infineon has implemented a Ďfab-liteí strategy, which boils down to the fact that it has stopped investing in certain parts of the value chain. Does this release financial resources that could now be used for other R&D topics?
Eul: We will keep our R&D budget relatively constant at about 17 to 19 percent of our revenues. The funds released by the fab-lite strategy are investments in the first instance. In this regard, Infineon has said it would reduce the investment quota for new manufacturing lines from the present double-digit percentage level to a single-digit percentage.
EE Times: Given the change in the value chain, which R&D topics do you regard as urgent?
Eul: For us as a company that sells complete system solutions, it is extremely important to have the expertise that enables us to know how an appliance will work and what it will look like. We regard it as one of our core competencies to understand all aspects of a systemís functionality and integrate it on to a chip. We spend our R&D bucks to enable us to offer complete system solutions to our customers.
EE Times: Does your fab-lite strategy imply a need to emphasize design-for manufacturing (DFM) activities?
Eul: No. DFM is important whether you manufacture in your own fab or in a foundry.
EE Times: In sophisticated microelectronic devices software is becoming increasingly important. However, software has not been a core competency for semiconductor manufacturers. How does a company such as Infineon acquire this competency?
Eul: Infineon has always pioneered the development of software for these kinds of chips. We do not need to acquire this competency, since weíve had it in-house for years. Someone who intends to sell system solutions needs to provide the software to run the chips. This is a significant factor for our competitiveness.
EE Times: Does this also apply for your activities in the automotive segment?
Eul: In this market, Infineon closely cooperates with system vendors such as Bosch, Continental, Delphi or Siemens VDO and defines the right combination for hardware and software in line with market needs.
EE Times: The fab-lite model could hit its limits when your manufacturing partners begin to build up their own design competencies and eventually start to compete with their customers. Do you see any danger of a conflict of interests?
Eul: This is theoretically possible. But reality shows that design and development costs are also increasing for foundries. Presently, we do not see any company that could succeed in breaking out of its own part of the value chain.
EE Times: Has Infineon implemented a key account strategy similar to the one in place at STMicroelectronics?
Eul: I do not know the key account strategy of our competitor STMicro-
electronics. We develop all our products in close collaboration with our customers. These customers are called key account here as well - companies such as Siemens, Ericsson, Nokia, LG, Panasonic, and also Huawei or ZTE. These companies are the key to our future - and thatís why we call them key accounts.
EE Times: In Japan, there are already concrete approaches to implement 4G mobile services with data rates in the 100-Mbit/second range. In Europe, one does not see much of this. Is the European mobile radio industry missing out on 4G development?
Eul: Certainly not. Japan is a country where new technologies are always accepted and deployed rapidly. This was very similar during the deployment phase of 3G mobile services. But nobody would say that Europe has fallen behind. For 4G, there are massive activities in Europe under way - mostly connected to LTE (Long Term Evolution).
EE Times: In the automotive electronics market, Infineon has a strong position; the same applies to power electronics. Against the background of the debate on climate change, this could lead to interesting synergies. In addition, Infineon has announced it plans to strengthen its position in the Japanese market which happens to be presently the only place where hybrid drive cars are built. Could you give us an update on your efforts with respect to Japan?
Eul: The situation is exactly as you have described. Indeed, at the time of the Qimonda spinoff, we announced that a part of the proceeds could be used to strengthen our automotive and/or communications activities. Also, we did not rule out acquisitions in Japan. But there is no news I can communicate.
Prof. Dr. Hermann Eul
Born: Mar. 1, 1959; Neustadt/Wied, Germany
GM Siemens Digital Telecom and Siemens Data Com ICs (1991-1999)
Infineon VP/GM Wireless Baseband and Systems, 1999-2001
Infineon, CEO Security & Chip Card IC business, 2001-2002
University of Hannover, Institute of RF technology and Radio systems: Professor and General Manager 2003-2004
Infineon Communications Solutions, SVP and GM 2004-2005
Infineon AG, Member of the board, overseeing Communications Solutions since July 2005
University of Applied Sciences Koblenz, Dipl.-Ing (FH) (comparable to Bachelor)1984;
Bochum University, Dipl.Ing (comparable to master of science/technology) 1987;
Bochum University Dr.-Ing (comparable to PhD/technology) 1990
This story appeared in the EE Times Europe print edition covering June 18 - July 1. European residents who wish to receive regular copies of EE Times Europe, subscribe here.
See other stories from this issue here.