MUNICH, Germany — Landshut Silicon Foundry GmbH (LFoundry) has launched its operations amidst what is possibly the worst crisis in semiconductor history. The startup is the result of a management-buyout of a Renesas production line. While some observers state the niche foundry market is already overcrowded and consolidation is in evitable, particularly in Europe, LFoundry co-founder and managing director Michael Lehnert, sketches out a bright future.

EE Times: Many chip companies, fabless and manufacturers, have been saying for a while that market visibility is poor. Could you help us to get a better perspective?

Michael Lehnert: Everybody, including the chip buyers, is reluctant to deliver any prediction. Our customers in this sense are no exception. We don't have much reliable data as to how and when an uptrend will emerge. In addition, we depend to a great deal on Renesas, our former owner and still our main customer for 2009. In the following years we will add more customers to our list. Thus, customer input for the current year is very limited. What I can say in this environment is that the market has clearly hit the bottom. We see a slight upturn in order entry for the Q2 and in particular for Q3.

EE Times: Well perhaps we should say that, for you, the visibility is good because you have this main customer that guarantees a certain level of utilization?

Lehnert: I would not overvalue that since this customer has the option to shift manufacturing orders between us, other foundries and internal production lines. We benefit from our high quality and delivery performance which is the reason why we have been able to pick up several additional orders. This does not necessarily mean that the market demand is already up.

EE Times: The large Asian foundries have reported improvements in their manufacturing capacity utilization recently. Is this also true for smaller European foundry service vendors?

Lehnert: Many of our customers are still in the prototyping phase or in the ramp-up process. For this reason, relevant statements from our customers refer to the year 2010 in which they certainly expect a recovery. For the current year I cannot say much since the volumes are still very small and the products have not yet reached their full production status. We only started to offer foundry services last year.

EE Times: Is the current slight upturn you have mentioned due to the fact that customers have to restock, but after having filled up their stocks the demand will continue to be at the same low level as before?

Lehnert: Yes, this is absolutely possible and we watch this process with a certain caution. We definitely do not believe that the industry in the next year will go back to normal as if nothing had happened. The growth ahead is lower than before and everybody is careful not to order more quantities than necessary to restock. Yes, the stocks at our customers are more or less empty and every order for them immediately triggers orders to the chip manufacturers. These orders currently are completed, but the industry is far from its earlier high levels. We tend to act very cautious at the moment.

EE Times: Low utilization at the leading-edge foundries brings them under severe price pressure. Do you see similar price dynamics acting in the analog/mixed signal foundry business?

Lehnert: Not quite. In comparison to digital foundries, in our industry there are very specific and different customer requirements. In 6- and 8-inch foundries * we are producing on 8-inch wafers * there is a lot of specific expertise and greatly varying process technology. Every foundry has different recipes [preventing easy movement of contracts]. In addition, business is designed for longer periods. We, for instance, have set prices in long-term contracts. For this reason, we do not observe such a price pressure. If there is price pressure in the analog/mixed signal foundry market, it is triggered by the transition from 6- to 8-inch wafers. This means that older technologies are more exposed to price pressure. For LFoundry with feature sizes of 0.35 through 0.15-micron the price pressure is limited.

EE Times: There is another dynamic in the market - in favor of the large and advanced foundries as IDMs are going fab-lite and outsourcing production. Do companies like yours benefit from this drive.

Lehnert: Recent market studies predict that in geometries below 65-nm there will be extreme growth over the next few years. LFoundry won't be able to benefit from this trend since our processes and equipment are laid out for 0.13-micron and above. We do not serve the market segment for 65-nm and lower, which for instance GlobalFoundries in Dresden is aiming at. There is a trend that analog will move to 200-mm wafers with geometries of 180-nm and 130-nm. This is what LFoundry will be able to benefit from.

EE Times: Do IDMs who go fab-lite compete with you? They typically are sitting on manufacturing capacities they need to fill. Taking orders from third parties could be an option which in turn would lead to increasing competition pressure for companies such as LFoundry.

Lehnert: The question is if these production lines can be maintained profitably. The IDMs would have to in-source orders from foundries to reach that goal. LFoundry however earns its money not so much from outsourced IDM manufacturing assignments but from fabless companies. We here have no other customer who could turn to such a strategy besides Renesas. However, I tend to believe that those IDMs will be forced to consolidate these older production lines as soon as they are no longer able to fully utilize their capacities. As a consequence, this effect can generate additional demand for foundries.

EE Times: Where will future growth be generated? Which industries, which applications will generate the demand to fill your production lines?

Lehnert: We have conducted detailed market studies, and from this it is clear that there are emerging markets with high analog content, particularly in the wireless domain. This is typically categorized under the "More Than Moore heading. Examples of this rapidly developing market include Bluetooth, Zigbee, Wifi, WiMax, NFC and UWB. Projections show these areas will demonstrate very strong growth over the next four years. Using an example UWB, a license-free technology group with high growth potential, the silicon real-estate for a chipset is estimated to be around 32 square millimeters. From this we calculate that the world market in 2008 equates to approximately 13,000 200-mm wafer equivalents. By 2012, this figure could climb to around 400,000 200-mm wafer equivalents. Market drivers include the rapid growth of wireless audio/video streaming. Factored into these calculations are a silicon real-estate shrink to 26 square millimeters through a combination of technology shrink to below 0.13 micron and design improvements. A similar growth potential is expected for ZigBee and WiMax, where we expect to see a ten-fold increase in volume over the next four years. Despite Wi-Fi being a fairly well established market, it is still expected to triple due to increased distribution and new technology features.

EE Times: Given the fact that your former mother company Renesas will gradually reduce orders, you must implement an expansive strategy. What do you believe is your unique selling proposition against the entrenched foundry competitors?

Lehnert: Our business model is to be the most advanced and flexible European analog mixed-signal foundry. Our basic ethos is to develop a lifetime partnership with our customers. With manufacturing in Europe, we are close to our main customers and LFoundry can offer advantageous pricing. Strong demand is being generated by the increase in bandwidth required by communications applications. This translates to larger, more complex designs, requiring more silicon real-estate. The digital content of these designs is relatively small, thus the ability to shrink is limited. However, significant gains can be made by taking advantage of advanced technologies to combine multi-chip designs into single chip solutions. For this reason, the combination of analog, digital and high voltage circuitry, in particular in automotive markets, will call for feature sizes to shrink from 180-nm to 110-nm over the next five years. Given these projections, we will be able to offer additional added value to our customers by manufacturing on more cost-effective 8-inch wafers in place of 6-inch. With the Landshut facility fully amortized within two years, the argument for 200-mm becomes even stronger, as the fixed costs for a 200-mm facility against those of a 6-inch facility are similar, while producing nearly twice the silicon real-estate.

EE Times: Are there technology features in this geometry segments that can constitute a competitive edge?

Lehnert: We are currently developing our own analog IP portfolio. For an analog/mixed signal foundry it is essential to offer an own differentiating IP portfolio since this will help to increase customer loyalty. This IP can be customer-specific, it can be developed in collaboration with the customer and with respect to requirements.

EE Times: Independently of this merger: If one thinks beyond the current crisis, how would you assess your perspective?

Lehnert: We will establish a customer partnership relation, generating customer-specific IP in growth markets. We are convinced that our offering will meet strong demand, and this is what we are already experiencing. We expect for the years around 2012 and 2013 an enormous growth, driven by an innovation thrust resulting from the integration of digital, analog and high voltage circuits as single chip solutions. With our technology we will gain a large share of the market.

EE Times: Does Europe as a location for semiconductor manufacturing have a future?

Lehnert: Clearly yes. Manufacturing in Europe is important since the innovations in electronics industry in and in the [automotive] supply chain will continue to be driven by progress in the semiconductor industry. If this innovation power should be maintained in Europe, it will be indispensable to retain the possibility for a close collaboration between designers, developers, users and of course manufacturing locally. In particular in Germany where innovation is generated to a large extend by small and medium-sized enterprises it is necessary to have a cooperation partner in the manufacturing industry. I even think there is space enough for two to three European foundries.

This story appeared in the May 2009 print edition of EE Times EuropeEuropean residents who wish to receive regular copies of EE Times Europe, subscribe here.

See other stories from this issue here.

You can download a digital edition of the latest EE Times Europe print edition here.