In an interview with EE Times Europe, X-Fab Silicon Foundries co-CEO Rudi De Winter reflects on company milestones such as its acquisition of 1st Silicon in Malaysia and its shift to manufacturing on 200mm wafers in Europe. De Winter also touched upon competitive and regional issues affecting X-Fab.

Germany-based X-Fab Silicon Foundries is doing well according to co-CEO Rudi De Winter, but as a private company it is not obliged to be very forthcoming with its numbers. The company is coming towards the end of a transition to 200mm wafer production and sees scope for expansion in its current fabs and in particular at X-Fab Sarawak, its wafer fab in Malaysia.

X-Fab, best known as a European foundry focused on analogue, mixed-signal and speciality processes, emerged from VEB Mikroelektronik in Erfurt, East Germany in 1992 soon after the re-unification of the country. The company acquired a few amortized, older wafer fabs and deployed a foundry service business model at a time when it was relatively new and mainly being applied to digital CMOS production.

The company, now majority owned by Belgium-based Xtrion, is focused on analogue and mixed-signal processes and with much of its work in the industrial and automotive sectors the company was not as susceptible to the global economic malaise of 2008 and 2009 as some other companies.

"We're doing well. We've grown through acquisition. Our latest one was 1st Silicon in Malaysia. That was fully loaded when we acquired it but we had to find new business," De Winter said. The acquisition took place in 2006 and part of the transition was to qualify the fab for the production of automotive ICs with multiple customers. "The fab runs 200mm wafers down to 130nm. It's the next generation from what we have in Erfurt," De Winter said.

De Winter said that in 2013 the Malaysia fab is at breakeven and that from now on he expects production there to grow profitably. "Manufacturing capacity there is about 20,000 wafer starts per month. We can further increase that to 30,000 wafers per month with additional equipment," said De Winter. "We installed 0.35µm HV CMOS applicable to automotive applications with 100V and a lot of different modules. We also have 0.18µm installed there."

Indeed overall X-Fab's highest shipping node is 0.35µm with mixed-signal capability, De Winter said. "The 0.35µm is still the most popular although 0.18µm is on a par with it in terms of enquiries."

X-Fab does have a more recent manufacturing acquisition than its fab in Malaysia. In November 2012 X-Fab increased its shareholding in MEMS Foundry Itzehoe to 51 per cent and renamed the company X-Fab MEMS Foundry Itzehoe. "Ah yes, but Itzehoe is MEMS only. The processes are varied so it is very hard to talk about wafer starts per month." But Itzehoe runs 200mm wafers as standard something that is unusual in the MEMS sector and which gives X-Fab an economic advantage.

The majority of X-Fabs' five wafer fabs are on 200mm. Erfurt runs a mix of 6-inch and 200mm wafers and Dresden will have converted to 200mm wafers only by mid-2014, said De Winter. The exception is X-Fab's facility in Lubbock, Texas, which runs speciality processes on 6-inch wafers. "In fact, we announced a deal with Cymbet so we are making batteries on silicon there," he said.

X-Fab has evolved in the analogue and mixed-signal foundry space by moving from 6-inch to 200mm diameter wafers and by moving into innovative processes such as MEMS and battery-on-silion deposition. But increasingly some of the digital giants of the foundry sector – TSMC, UMC, Globalfoundries – are also offering 'more-than-Moore' processes using their own older wafer fabs. How does X-Fab cope with that level of competition?

"If you look across the total foundry landscape there are many processes but they are not all offered in all market sectors," he said. X-Fab is not playing in the pure digital sector nor does it require the very high volume orders that are normal for the larger foundries. "When it comes to more-than-Moore when you look at the levels of integration for non-volatile memory, for high voltage, across a broad temperature range, I believe we are un-matched," said De Winter. "We also put a lot of emphasis on design support, we invest a lot in models, in PDKs [physical design kits] to give customers the best chance of success. Our NVM IP is developed in house, which is important for the automotive market."

De Winter said that in the automotive sector multi-party supply chains are not favoured by customers that must control and attest to the rigor of design and manufacturing for safety.

Emphasizing the support to design teams that X-Fab can provide De Winter added: "In MEMS we also have IP blocks and offer pressure and inertial sensor platforms."

X-Fab's sales by major industry sector breakdowns as: automotive 46 per cent, industrial 15 per cent and communications and consumer about 33 per cent. The high automotive tally is largely due to the work done for automotive component company Melexis NV, a sister company that is also controlled by Xtrion. "Melexis is around a third of our business," said De Winter.

Does X-Fab's acquisition of 1st Silicon as the next domino in its chain of globalisation indicate that is it grows it is choosing to, or being forced to, de-emphasise Europe? "We are glad to have our fab in Asia. It is well received by our Asian customers," De Winter responded.

But European Commissioner Neelie Kroes with her call to create an Airbus of Chips and her 10/100/20 strategic plan wants $13.5 billion of European tax-payers' money to leverage €100 billion of industry investment to enable Europe to double manufacturing as percentage of the global total to 20 per cent by 2020. Surely that represents a regional opportunity for X-Fab?

"I am glad such statements are made, but 20 per cent is a very aggressive number. The European Commission wants to stimulate the semiconductor equipment industry for 450mm diameter wafers. That's not something for us. They also want to stimulate 'more-Moore' [leading edge], which is for STMicroelectronics and Globalfoundries. More-than-Moore is somewhere we could participate and we want to contribute and also to grow. We are working in MEMS in Europe for example. But it is not yet clear how this strategy will benefit us."

De Winter is also sceptical about plans for setting down large fabs for More-than-Moore operations such as MEMS. "If you look at the MEMS market in mobile the volumes are big but the margins are small. There are a lot of smaller MEMS opportunities in market niches. In those sectors it is often engineering and research that makes the difference rather than just manufacturing capacity. And, in fact, Europe is well placed in that regard."

"We will be acquiring one day. We get offers of fabs to buy every day but there is nothing very serious at the moment. At present we are consolidating in our current facilities. It might be something to think about in three-years' time."

And what about an IPO for the privately-held parent Xtrion NV or X-Fab to fund expansion? De Winter indicated that this was unlikely on the grounds that if the business model is working why try to fix it.

By: DocMemory
Copyright © 2023 CST, Inc. All Rights Reserved