Skip to main content



(INTC) - Get Intel Corporation Report

developers conference kicking off Tuesday, expect to hear lots of talk on the wireless front.

Case in point: Intel has been chatting up its plans to squeeze more silicon functions on a single piece of silicon for cell phones, a product code-named Manitoba.

It's part of an evolution toward increasingly integrated semiconductors. But while consumers benefit as ever-smarter and cheaper devices become available, the takeaway for chipmakers is less benign.

Intel's entry into the wireless space is bound to escalate pressure on chip prices for existing players, though the company probably won't be a big presence for another couple of years. Meanwhile, advances in integration demand that chipmakers make further R&D investments at a time many would like to trim their operating expenses, after heavy spending in the late '90s.

The (dubious) prize for Intel is a chance to gain share in the handset market. To chipmakers demoralized by the torpid PC market, cellular handsets offer at least the potential for unit growth. Sales of handsets are expected to jump from around 400 million last year to around 1 billion by 2007 or 2008. It's a fresh market for Intel, whose wireless revenue until now has come almost entirely from flash memory chips.

But though handset sales will rise, increasing competition means the cost of the chips that go into them probably will stay under pressure. Last week, rival

Texas Instruments

(TXN) - Get Texas Instruments Incorporated Report

reminded the market that it aims to debut the cell-phone version of its system on a chip in 2004. It's currently shopping around a Bluetooth version of the chip.

"I believe as everybody tries to drive down costs per unit and chase units and the revenue opportunities associated with them, the likely result is continued margin pressure," says Adam Parker, an analyst at Sanford Bernstein.

From that standpoint, chip integration is great news for cell-phone consumers. "These components are roughly one-third of the cost of the whole handset," explains Will Strauss, president of Forward Concepts, a market research firm. "The more components you can get on a chip, the cheaper and smaller a cell phone gets."

TheStreet Recommends

In a sense, leading chipmakers like Intel, TI,



ST Micro

(STM) - Get STMicroelectronics NV Report

don't have any choice but to pursue integrated chips, since it makes their products less expensive and easier to market. "The pricing pressure on chipmakers is coming from the OEMs and ODMs building the phones," who want ever-cheaper components, explains Allen Leibovitch, an analyst at IDC. "

Integration is the response to the pricing pressure."

It also offers potential for growth in an industry in which it's been sorely lacking. "Companies that can't integrate risk losing significant revenue opportunities," says Parker. "But the very process of integration is capital intensive and R&D intensive, and it may have the effect of suppressing margins for the winners."

Intel is sampling the chip to vendors now, but hasn't yet announced any design wins. More details are likely to emerge from the developers forum this week. For now, the announcement of the chip, while expected, has led to speculation about the market-share fallout.

"The big thing here to me is that Intel is getting a slice of the market they haven't been in before, going after the digital baseband business," says Tim Shelton, a senior analyst at Allied Business Intelligence. "TI, Qualcomm, ST Micro -- those types of suppliers are certainly paying attention to what Intel has just announced."

In an interview, Dennis Buss, vice president of silicon technology development for Texas Instruments, acknowledged Intel as a "fearsome competitor" at the same time he touted the benefits of TI's own integrated cell-phone chip, which won't be available until next year.

Still, Intel's offering is far from a shoo-in. For example, rivals such as TI offer wireless chips with less performance-intensive processors. "They're cheaper and smaller than an Xscale," Intel's version, explains Leibovitch. "So Intel's definitely aiming at the higher end. Demand could be an issue and the price point could be an issue."

"TI has a lot of design wins with a lot of phone companies," says Shelton. "Intel has to catch up."

Strauss doesn't believe Intel will be able to ship Manitoba chips in volume this year. "The most they could hope for next year in their wildest dreams is around 10% of the market." But assuming handsets grow to around 500 million units by then, that means Intel could sell 50 million chips, he notes. "That ain't bad -- they're shipping about 100 million PC chips right now," he says. "Admittedly, these are smaller and cheaper than Pentium, but they'll have to learn to live with less."