Page 4

EETE JAN 2014

What a mess! By Peter Clarke The closure of LFoundry Rousset, ordered by the Commercial Court in Paris, puts an end – probably permanently – to chip manufacturing in Aix-en-Provence; a blow to France, to Europe and to 613 engineering workers that were employed there. We even have the parent company, LFoundry in Germany, reportedly under a criminal investigation in Paris for the alleged misappropriation of funds from LFoundry in France during dealings prior to this unhappy state of affairs. The fighting and dissent that have now broken out within Europe are all so different to the expansive, co-operative, pan- European days that created the Rousset fab in 1985. Clearly one of Rousset’s problems is that it runs 200mm wafers – when much manufacturing has moved on to larger, more economic 300mm wafers – and it has relatively mature manufacturing processes in place – although not disastrously so for mixed-signal IC manufacturing. It has a manufacturing capacity of about 25,000 wafer starts per month, and it is – or was – a repository of skills that could have been useful to wealth creation in Europe in the decades to come. And yet only two years ago Rousset was being positioned as a beacon of hope within LFoundry. LFoundry GmbH was founded in 2008 by the management buy-out of a Renesas wafer fab in Landshut, Germany. The purchase came with legacy orders from the vendor but LFoundry put the German manufacturing site out into a separate entity called Landshut Silicon Foundry GmbH, which was declared insolvent in 2011. And this was because no sales were found to replace those legacy orders from Renesas. Meanwhile in 2010 LFoundry had acquired the Rousset fab from Atmel Corp. together with legacy orders. But again LFoundry has failed to replace those legacy orders to a significant “Only two years ago Rousset was being positioned as a beacon of hope within LFoundry” degree. It is reported that Atmel has contributed to Rousset’s downfall by cutting back orders after three years. But it is any foundry’s job to make its chip manufacturing attractive so that customers are lining up, not turning away. And so Rousset is the second LFoundry wafer fab that has had to close due to being unable to make sales after legacy chip orders have gone away. And even as Rousset closes LFoundry is in the process of bringing up its third fab, a former Texas Instruments and then Micron Technology wafer fab in Avezzano, Italy. It is as if LFoundry is using legacy fabs as stepping-stones to make its way across Europe – from Germany to France to Italy. On present form the Avezzano wafer fab has three years before its demise. The more cynical amongst us might see a broader and more general pattern at multiple management buyouts across Europe whereby inward investors are relieved of their responsibilities to long-serving employees by transferring ownership to a new company for the last couple of years of a factory’s life. Standing back it can be seen that Europe has been managing – or mismanaging depending on your point of view – its exit from the semiconductor industry for some time. And in the particular case of Lfoundry there seems to have been a decided lack of ability to create new business. And yet it need not be so. There are other companies that have made a good job of taking on legacy manufacturing and creating new business. There’s X-Fab Silicon Foundries AG in Germany and Tower Semiconductor in Israel. Both are expansive and global in their ambitions. Tower has successfully taken on a Micron wafer fab in Nishiwaki, Japan, and has struck a deal with Panasonic Corp. to take control of three of the Japanese company’s wafer fabs including 300-mm capable fabs. It is all quite different to 1985 when Rousset was the manufacturing site for a new company, European Silicon Structures (ES2). ES2 was funded by European industrial giants of the time, that were users of ICs and just getting into custom IC design. They included Philips Electronics, British Aerospace; Olivetti; Sweden’s Saab France’s Bull, Switzerland’s Asea Brown Boveri and Societe Generale de Belgique. The whole point of the company was to apply Silicon Valley business methods learned from inward investors in chip making, such as Motorola and National Semiconductor and let a European company apply them, grab European market share and eventually take global market share. One wrinkle in the business plan that would eventually prove ES2’s undoing was a decision to use direct-write e-beam manufacturing in an attempt to create low volume chips while saving on mask costs. The plan was technically feasible and differentiating, but it put ES2 outside the mainstream of chip manufacturing and in a niche between FPGAs and the semi-custom and full-custom offerings of LSI Logic Corp., VLSI Technology Inc. and 4 Electronic Engineering Times Europe January 2014 www.electronics-eetimes.com


EETE JAN 2014
To see the actual publication please follow the link above