), whose strongholds are chips used in communications and consumer electronics.
What's so attractive to Intel (INTC
)? Consider that demand for semiconductors used in flat-panel TVs has been doubling each year. Or take chips used in what's becoming one of the world's most ubiquitous products: the cell phone. Demand there is scintillating, with sales of key DSP (digital signal processor) chips rising 28% last year, to $6.2 billion, and expected to reach $9.6 billion in 2006, according to the Semiconductor Industry Assn. By comparison, sales of Intel's bread-and-butter PC processors should rise by a comparatively modest 14.5%, according to tech consultancy IDC.
Erach Desai, an analyst with American Technology Research in Boston, goes so far as to wonder if TI is the ascendant company: "Intel has been the bellwether for the semiconductor market for two decades," he says. "Now, perhaps TI, because of the markets it plays in, is better" for investors looking for faster growth.
DOUBT AND DOOM? No wonder Intel is redoubling its efforts to expand into mobile phones and consumer electronics. On Apr. 12, it announced new cell-phone chips that will compete directly with TI's application processors, responsible for handling a phone's software. And in February, Intel announced another competitor to TI: a baseband chip, which takes care of the communications functions in some handsets.
Those came on the heels of a January announcement that Intel is developing an alternative to TI's DLP (digital light processor) chips, used to project images in flat-panel TVs. DLPs were among the biggest contributors to TI's growth in the latest quarter, with their sales rising more than 50% sequentially, according to the company.
Intel's moves have led to a media frenzy and much speculation of doubt and even doom for the likes of TI (not to mention other players in these markets, including Hitachi (HIT
), Philips (PHG
), and STMicroelectronics (STM
). But those scenarios might be a bit premature. Intel's product portfolios are far from complete. And these markets will be hard to tackle and will take time to penetrate. Even Intel admits as much, says David Rogers, its wireless marketing manager.
NO PARTNERS. To start with, the technology Intel is developing for flat-panel TVs is hardly ready to start showing off in living rooms. Sure, Intel has demoed its so-called LCOS (liquid crystal on silicon) technology for image projection, which most analysts agree has major cost advantages over DLP. But Intel has yet to make any chip product announcements -- nor has it signed up any TV makers as partners. And rivals, such as Toshiba (TOSBF
), which have also tried to develop LCOS chips, have had troubles achieving the image quality and reliability TVs require, says Bert McComas, an analyst at market consultancy DisplaySearch in Phoenix.
Indeed, analysts such as Tai Nguyen of Susquehanna Financial Group in Bala Cynwyd, Pa., don't expect Intel to see any significant revenue from LCOS chips for the next two years. And it might not have even a 10% market share in three to four years, McComas says.
In its cell-phone effort, Intel's upcoming baseband chip addresses a tiny market. Out of 570 million handsets expected to be shipped this year, only about 10 million will feature the wireless technology called WCDMA (wideband code division multiple access) that this chip uses. WCDMA enables advanced functions such as video-streaming to a phone, says Allen Nogee, an analyst with tech consultancy Cahners In-Stat.
He blames the slow adoption of these phones on their short battery life and high price. And most of the new phones will be sold in Japan, where Intel competitors NEC (NIPNY
) and Matsushita (MC
), not to mention market leader TI, already have a strong foothold.MISSING PACKAGE. Even closer to its computer-processor home turf, Intel's new products face tough going. The original Xscale chip, used in personal digital assistants made by the likes of Hewlett-Packard (HPQ
) and Samsung, grabbed a hefty share of the PDA market, but to date they've appeared in only about half a dozen cell phones. So, Intel now has a new Xscale chip family, called Bulverde (after a city in Texas), which offers more powerful chips. But they may run into the same problems as their predecessor: Intel still offers a much more limited selection of chips than its rivals, and most cell-phone makers prefer to buy a complete package of chips needed for a phone -- something TI does offer.
In addition, the cell-phone market is tricky to penetrate, thanks to existing relationships between phone makers and chip suppliers. Market leader Nokia (NOK
), for one, uses TI's software and chip designs in its phones -- and redoing all that would take months, if not years. And given the intensifying competition from Chinese and Korean phone makers, Nokia is unlikely to make such a switch any time soon.
TI even keeps its inventory at its largest customers' plants, so the sale essentially occurs when the customer takes out that inventory for production. Intel could conceivably offer that, too, but making a profit on this setup requires major economies of scale.
FAB EDGE. Moreover, Intel's credibility with cell-phone makers was hurt last year when the company increased prices on its flash memory -- and, as a result, lost its leading market share to Samsung, says Will Strauss, president of tech consultancy Forward Concepts. Finally, Intel has lots of distractions in its core business, where it faces stiff competition from its main processor rival AMD (AMD
), whose latest PC chips have rated better in certain tests.
That doesn't mean Intel can't ultimately succeed in these markets, but it does mean it would take time. Intel is still leading the industry in cutting-edge chip manufacturing and, unlike many rivals, it owns its own chipmaking fabs. That allows for great economies of scale and lower production costs.
Intel's long-standing relationship with software powerhouse Microsoft (MSFT
) should also help. As versions of the Windows operating system and other Microsoft software make their way into advanced cell phones, Intel's chips should penetrate those markets, too, says Nguyen. However, building up the software support might take a few years.
NOT THERE YET. Also, Intel is slowly but surely developing relationships with wireless carriers -- which increasingly have more say over what cell phones look like. In March, Intel agreed to market and sell Sprint's (FON
) wireless services with products -- such as notebooks that can access the Web through Sprint's networks -- containing Intel processors. Such relationships could, perhaps, extend in the future to include phones.
Still, before Intel gets to that promised land of being the chipmaker for every device, it'll have to prove it can oust TI from some choice real estate first. By Olga Kharif in Portland, Ore.