Time For Infineon To Forget Memory Chips?

Infineon, lacking the scale and lower costs of its rivals, may tighten its focus

Stay the course? Or cut and run? That seems to be the choice facing beleaguered German chipmaker Infineon Technologies (IFX ). A late November board meeting could set the stage for Infineon's retreat from the memory chip business.

It has already been a tumultuous 18 months for Infineon. Last year the board ousted flamboyant Chief Executive Ulrich Schumacher amid disagreements over strategy and his overbearing management style. His successor Wolfgang Ziebart, a veteran of German tiremaker Continental (CTTAY ), has shaved some $236 million off Infineon's fixed costs this year.

But Ziebart hasn't been able to change the brutal dynamics driving the chip business. Infineon, squeezed by Asian competitors, simply may not have the staying power to keep competing in basic memory chips, which generate 40% of its sales. The company is expected to post a loss of about $267 million for the fiscal year ended September 30, on sales of $8 billion. "This really is a scale business, where only the big can thrive. And at present, Infineon just hasn't got enough scale to guarantee profitability across the cycle," says Uche Orji, a London-based analyst at JPMorgan Chase & Co. (JPM ).

Chipmakers in this segment are under constant pressure, with the price of DRAM memory chips falling by an average of 32% a year, according to industry researcher Future Horizons. In addition, the business requires multibillion-dollar investments by anyone wanting to stay ahead in the technology. DRAMs can be hugely profitable, but difficult years such as 2001 and 2002 helped drag Infineon into billion-dollar losses.


To compensate, the company has expanded its business in specialty memory chips -- faster and higher-margin chips used in game consoles, digital cameras, and other devices. In August, Infineon won a contract to supply chips for Microsoft Corp.'s (MSFT ) new Xbox 360 game console. Trouble is, specialty chips account for only 6% of Infineon's memory product sales, vs. up to 30% at Samsung, according to HVB analyst Günther Hollfelder.

Investors now think Ziebart and his team are looking to exit the memory products business. Such a move would transform Infineon into a much smaller manufacturer of logic chips for automotive and telecom clients. Infineon is not saying what it may do -- but it has indicated that its strategy review later this month will be more intense than usual. The betting is that after the board meeting on Nov. 17, Infineon will at least clarify if it will begin the search for a partner or buyer for its memory business. The stock will doubtless come under pressure if no announcement comes.

But even if Infineon decides to get out of memory, it cannot simply sell off or float the unit. High production costs -- key sites are located in high-cost Germany and the U.S. -- make the business expensive and unattractive to most investors. "Face it, who would want a company with a European cost base and poor profitability?" says one analyst.

Industry watchers speculate that an expanded partnership with Taiwan's Nanya Technology Corp., one of the few companies that shares Infineon's production technology, is the most feasible move. Merging their DRAM operations would bring crucial scale while sharing out the huge investment costs. But the companies are likely to disagree widely on how the business should be valued. Taxes and intellectual-property rights would also complicate any deal, says Andrew Griffin, an analyst at Merrill Lynch & Co. (MER ).

Finally, even assuming that Ziebart can manage an exit, he will still have plenty of work left to do. Infineon's "crown jewel" -- its chip business for the auto industry -- is profitable and growing. But its business for fixed-line telephony is ailing, while chips for mobile phones came under pressure when handset sales at Infineon's key client, Siemens (SI ), fell sharply.

Infineon may yet decide to keep betting on DRAMs: Its other units benefit from the cutting-edge technology developed in that division. The company can stand to support its commodity memory business a while longer, since the next round of investment in memory technology is years away. But Infineon cannot delay a decision indefinitely. "If the company does nothing, it will gradually bleed to death," says Orji. Not a pretty scenario.

By Bettina Wassener

    Before it's here, it's on the Bloomberg Terminal.