SAN FRANCISCO -- Spansion ( SPSN) has long lived in the low-rent district of the flash memory market, making chips for applications with limited growth prospects and sex appeal. Now the chipmaker believes it has the secret sauce to take over the entire flash market. At first glance, Spansion's announcement Monday of its plans to acquire Israel's Saifun Semiconductors ( SFUN) for $368 million is a bit of a head-scratcher: Spansion currently licenses Saifun's technology. So the deal, while saving it money on licensing fees, essentially gives Spansion something it already had. And Saifun has historically had a
rough time trying to license its technology to other companies. If Spansion was looking for a way to generate new revenue through intellectual property licensing, it hasn't exactly picked a very hot commodity to peddle. Saifun's $8.7 million in sales in the most recently ended quarter were down 50% year over year -- after Infineon Technologies ( IFX) opted to cease licensing its technology -- and the company's shares are down 65% in the past 12 months. This may explain the negative reaction among Spansion shareholders. The stock slipped more than 3% on Monday, and was recently off another 5 cents to $8.04 in Tuesday trading. But executives at Spansion and Saifun are banking that the flash memory market is on the verge of a major change -- and that the two companies have a much better chance of coming out on top by joining forces.
"We see humongous opportunities for our companies and now is the time to position ourselves as the company that's going to be the solution for the future," said Spansion CEO Bertrand Cambou in a conference call Monday. Flash chips are a form of memory that can store data even when the power is switched off. Sunnyvale, Calif.-based Spansion, which lost $148 million on sales of $1.3 billion last year, has traditionally focused on a flavor of flash known as NOR, which is used to store the software that operates electronic devices. The biggest growth opportunity going forward, however, is in NAND flash, which is used to store the ever-expanding libraries of digital photos and music on gadgets like MP3 players, cell phones and digital cameras. The problem with NAND, however, is that it's based on a technology known as floating gate, which appears to be approaching a technological dead end. Many people believe floating-gate NAND will run into major manufacturing problems as the circuits in chips are shrunk to a certain point. Last month, Intel's ( INTC) storage technology and manufacturing director, Knut Grimsrud, told attendees at the company's developer conference that NAND flash should be able to scale to smaller-sized circuits through the end of the decade, but that new technologies might be required after that.
That's where Saifun sees its technology coming in. According to executives at the company, Saifun's technology doesn't suffer from any of the miniaturization problems found in floating-gate NAND. And Saifun's technology is suitable as the basis for all sorts of memory -- NAND, NOR, even DRAM. As NAND's expiration date on floating gate gets closer, Spansion believes companies will have no choice but to license the Saifun technology it owns. The change will occur as soon as the next 12 to 18 months, said Cambou. Of course, predicting an industry's move to a new technology standard is inherently tricky. Businesses have a way of figuring out clever workarounds that extend the status quo, rather than taking abrupt -- and risky -- jumps to entirely new technologies. But Jim Handy, a memory analyst at industry research firm Objective Analysis, contends that switching NAND manufacturing from today's technology to the Saifun technology is not as big a leap as it might seem. Saifun's so-called charge-trapping technology doesn't involve using any new processes or chemicals in a standard silicon manufacturing line, he notes. "There's no question that all flash manufacturers are going to use charge-trapping in the future," says Handy. "The question is whether or not Spansion can use Saifun's IP
intellectual property to get royalties from that."
Whether companies might be able to adopt a generic version of charge-trapping technology that doesn't require a license from Spansion remains to be seen. But Spansion's Cambou believes companies will take a license because it will get them up and running faster. Spansion has first-hand experience manufacturing the technology in volume and spends $450 million a year on R&D -- expertise and resources that Saifun could not bring to the table when it was trying to license the technology as a standalone company. "This is the case where we will have solutions and we can help them to save time," Cambou said in an interview. Spansion also will inherit Saifun's technology related to flash memory that stores four bits of data per cell, instead of today's two bit per cell maximum. Since it reduces costs, the technology could prove appealing when offered by Spansion, given that it's the only company to have successfully commercialized four-bit flash so far. Spansion executives hope to add new licensees next year, after the transaction closes, although the company doesn't expect the acquisition to hurt its finances even with the current level of licensees. The deal, said company officials, should expand profit margins and accelerate Spansion's path to profitability. Spansion's ambitions to conquer the flash world are hardly a sure thing, but the company may have found a relatively painless way to improve its standing in the flash world.