), the world's largest maker of microprocessors, has been trying to offset the effects of stagnating PC demand by expanding into the faster-growing market of chips for communications devices. It's already the world's No. 1 seller of flash memory, which is used to store data on mobile phones and other digital devices. This year, Intel is turning its communications strategy up a notch -- to focus even more on the market for other cell-phone chips. And that's about to start a war between it and the companies that now dominate that market, including Texas Instruments (TXN
Intel's entree into the market is the increasing sophistication of cell phones, which are starting to incorporate the capabilities that Intel's microprocessors bring to PCs. Intel's Manitoba product, announced in February, featured its first digital signal processor (DSP), a chip that converts analog signals, such as sounds, into digital form, plus a processor and flash memory -- all in one chip. Manitoba, which starts production in the third quarter, is Intel's way of showing that it can be a one-stop shop.
The plan is to eventually be a major supplier of everything inside a cell phone. Intel should grab a 20% market share of cell-phone silicon within five years, up from a single-digit share today, estimates Alan Nogee, an analyst with Cahners In-Stat.
EARLY FAUX PAS. That'll mean pushing its way into a market now held by TI, Qualcomm (QCOM
), Motorola (MOT
), and STMicroelectronics (STM
), all of which are gearing up to fend off Intel. The competition will burst into open conflict later in 2003, when these rivals unveil or start producing their newest offerings for the newest souped-up cell phones, called smartphones.
Uncharacteristically, Intel has made one mistake that could come back to haunt it. At the beginning of 2003, it raised prices on flash memory by 20% to 40% -- and in the first quarter lost 7% of its market share, down to 17%, to competitors such as Advanced Micro Devices (AMD
), estimates Alan Niebel, CEO of market researcher Web-Feet Research. Since then, it has regained some of that, Niebel adds.
However, its price increases offended the same cell-phone makers it needs to woo to make its broader communications strategy work, says Dan Niles, an analyst with Lehman Bros., whose compensation is not tied to the stock. "It shot itself in the foot," he says.
KOREAN THREAT. The price hike encouraged cell-phone makers to look for cheaper alternatives to Intel's flash-memory products. One thing they turned to was so-called NAND flash, says Tom Quinn, a vice-president at Samsung in North America (Samsung holds 55% market share in NAND chips). An Intel spokespers says NAND is less stable and won't be used in cell phones. But many analysts say otherwise. And the threat from Samsung will only increase as, later this year, the Korean manufacturer is expected to expand into other types of memory used in mobile phones, says Jim Cantore, an analyst with semiconductor consultancy iSuppli.
Intel's flash gaffe could add to the unease of some cell-phone manufacturers over the chip king's new strategy. They "wouldn't want Intel to standardize the cell-phone platform in the same way it standardized [the platform for] PCs," says Dan Scovel, an analyst with Needham & Co. After all, this could mean that a company like Nokia (NOK
), the world's largest maker of cell phones, would see its margins thin -- and go to Intel. Indeed, although Intel has sold Manitoba to Korean manufacturer Maxon and others, it has yet to announce a major customer for the chip. (Motorola will use a different Intel product in a phone due to come out this summer.)
That could partly reflect a concern by handset makers that Intel hasn't integrated all the right elements onto the chip. Cell-phone memory requirements are increasing rapidly, and a chip with integrated memory won't be able to keep pace, some analysts say. Intel expects Manitoba to do well in niche markets, according to its spokesperson. But unless the chipmaker comes out with more -- and better -- versions soon, it could find that gaining ground on rivals such as TI will be tough. Intel is planning to launch a more sophisticated successor to Manitoba in November.
"FORMIDABLE COMPETITOR." It also needs to further expand its product portfolio to include other cell-phone components, such as some radio parts responsible for phone transmissions. Others, including TI, already make those, notes Tim Shelton, an analyst with Allied Business Intelligence, a technology think tank in Oyster Bay, N.Y. In addition, Intel needs to make chips for more types of technologies used in wireless networks today, analysts say.
Yet, rivals are nervous. "Intel is a formidable competitor," says Loic Lietar, general manager of STMicro's wireless business unit. In early 2004, STMicro will start shipping its Nomadik chip, integrating a processor and a DSP -- the "most important product" now in its pipeline, says Lietar. Meanwhile, TI is testing several new products and is on track to combine all chips that go into a cell phone onto one piece of silicon by the end of 2004, says Avner Goren, who heads that effort. That would be an advance compared to Intel's Manitoba, which combines only a few of about a dozen chips that go into a cell phone.
All the chipmakers are betting on the evolution of cell phones into much more sophisticated devices. As that happens, Intel figures, handsets will need the powerful processors like those it makes for computers -- in a compact package that helps keep phones small and cheap. One key to its success will be aggressive research and development, funded by Intel's $9 billion in cash. For example, it's working on a DSP chip that consumes one one-hundredth or less of the energy that today's chips use. That would allow manufacturers' to increase the phones' battery life. "If you come up with any definition [of what the cell phone of the future would look like], we'll have the capability to make parts for it," declares Shekhar Borkar, director of Intel's circuit research.
Indeed, most analysts expect Intel to gain 20% of the cell-phone chip market within five years. That would mean significant new revenue for the $26.7 billion company, since the value of chip content in a typical cell phone already ranges from $50 to $200 -- and is increasing, says Nogee. And about 450 million new phones will be manufactured worldwide in 2003. But with everyone else gearing up for battle, Intel has little room for error. By Olga Kharif in Portland, Ore.