The acquisition of a 200mm wafer semiconductor plant from Motorola would help Semiconductor Manufacturing International Corp (SMIC) expand capacity and broaden its product portfolio beyond memory chips. It would also give Motorola the chance to divest itself of an unprofitable investment that has yielded little but headaches since the project was launched in 1995. For the past two months, reports have suggested SMIC is in discussions to swap shares for a stake in Motorola's factory, based in Tianjin, although both parties have declined to comment. A sale to SMIC would fit with Motorola's 'asset-light' strategy, whereby it leaves the manufacturing of its chips to others. Last month, Motorola chip chief executive Scott Anderson said he was unhappy with the Tianjin fabrication facility - known as a fab - and was looking for a partner for the operation. 'That fab just does not fit into what Motorola is trying to do,' Merrill Lynch chip analyst Dan Heyler said. Motorola bought land for the Tianjin plant in 1995 with the intention of producing chips to go into cars, but progress was delayed by an economic slump in Asia and a downturn in the semiconductor industry. Work began in earnest in 2001, with Motorola pouring roughly US$1 billion into the project. The initial hope was that a factory based in China would help Motorola gain access to the mainland market for its products and build relations with the central government. 'China wanted them to have local production,' Mr Heyler said. Earlier this year, Motorola said it would begin to increase production at the plant. The factory is capable of making about 24,000 wafers per month but production has been kept to an estimated 12,000 wafers monthly as demand in China turned out to be less than initially hoped. Turning the operation over to SMIC - which makes chips on a contract basis - would allow the plant to use more of its capacity and help it achieve profitability. The factory presently manufacturers chips solely for Motorola. 'The fab will be better utilised if more people can place orders with it,' a Hong Kong-based fund manager said. 'But in the end, I don't think we need all these fabs.' In 2001, mainland contract foundries announced plans to build as many as 25 200mm fabs, but several of these projects have been delayed or cancelled as capacity exceeded what domestic demand and overseas orders could support. For SMIC, making chips for Motorola and other customers at the Tianjin fab would help it diversify its product offerings. So far, the chipmaker's primary clients are memory manufacturers such as Germany's Infineon Technologies and Japan's Elpida Memory. The heavy emphasis on memory manufacturing leaves the company vulnerable to wild price swings in the volatile market. 'To date, no other pure play foundry has been successful using this model,' research firm iSuppli noted in a report. Motorola presently makes metal-oxide semiconductors and other chips at its Tianjin plant. The plant would complement SMIC's three 200mm lines in Shanghai and a 300mm plant under construction in Beijing. SMIC also has plans to build two more 200mm plants in the capital. But market watchers question whether the mainland needs the advanced 300mm fab, saying SMIC may not attract enough orders to make it profitable. The 300mm line, which would initially cost an estimated US$1 billion, would use technology supplied by Infineon to produce memory chips of a line-width as small as 0.11 micron. 'Memory doesn't have very high margins, so they have to be incredibly efficient,' iSuppli analyst Len Jelinek said. 'That thing is going to have to ramp and it's going to have to ramp hard.' He estimated it would take three years before the 300mm plant turned a profit. 'You won't see any revenue out of that facility until late next year [when it comes online]. That's just sunk costs.'