One of the hallmarks of Craig Barrett's reign as the chief executive of chipmaker Intel was his penchant for acquisitions in the area of communications.
In March, 2000, for instance, Intel (INTC) spent $1.25 billion to acquire Giga A/S, a Danish concern focused on building optical communications chips. The year before that it spent $2 billion and change on Level One Communications and another $1.6 billion on DSP Communications.
With PC sales growth slowing and networking and wireless telecommunications gaining momentum, it seemed the world's largest chipmaker could bring its considerable powers to bear, and in time dominate them all.
BACK TO BASICS. Things didn't quite turn out that way. Established players, such as Texas Instruments (TXN), doggedly refused to cede ground, and Intel found itself in the unlikely role of having difficulty landing customers. Meantime, Intel stuck to a bet on a particular kind of memory known as NOR flash, even as wireless companies favored a technically inferior -- though cheaper -- type called NAND championed by rivals. That contributed to the company's failure to create successful cellular platforms with its various types of chips, though it has won some high-profile wireless customers, including Research In Motion (RIMM).
Still, the net effect has been that some business units within Intel are losing money. And under CEO Paul Otellini, Intel is increasingly falling back on what it knows best: microprocessor chips for PCs and servers, chipsets and motherboards to accompany them, NOR flash memory chips, and its collection of chip factories around the world(see BusinessWeek.com, 5/23/06, "Intel Strikes Back,").
But as Otellini said at Intel's analyst meeting last month, every money-losing segment of Intel's business -- or as he put it "everything with a bracket around it" -- is under scrutiny (see BusinessWeek.com, 4/28/06, "Intel on the Offensive,". And that has touched off feverish speculation that a sell-off of certain Intel business units is imminent.
GUESSING GAME. Publicly, Intel has said very little about the matter, though Otellini has denied reports that the company's unprofitable NOR flash memory unit is on the block. A spokesman says the company is still in the middle of a 90-day review process and that an update on the results of that review isn't expected until July.
Analysts say they won't be surprised by big changes once the review is completed, and suggest that possible targets for sale or spin-off include Intel's wireless chip unit, which produces, among other things, the XScale line of chips used by RIM, maker of the BlackBerry wireless handheld device, and other outfits, including Palm (PALM) and Garmin (GRMN).
Analyst Doug Freedman of American Technology Research in San Francisco says selling the XScale business now might make sense. "This might turn out to be a compelling time to sell that unit," Freedman says. "It's not floundering as badly as it once was and is starting to generate some real interest. But it's not on fire," he adds. "I have to wonder why doesn't RIM just step up and buy it?"
HP HOVERING? Other potential buyers for the XScale unit could be Texas Instruments, the market leader in chips for wireless phones, Broadcom (BRCM), Qualcomm (QCOM), and Freescale (FSL)-- all players in the wireless chip business, says Instat/MDR analyst Jim McGregor. "I can't confirm that it's up for sale, but I know that the cellular chip group is under scrutiny," McGregor said.
Another likely target in a sell-off scenario, Freedman says, is the Itanium server chip unit. Intel and Hewlett-Packard (HPQ) together spent billions developing the Itanium chip. Though Itanium represented a fundamental rethinking in how chips used in high-end corporate computing are designed, rival AMD (AMD) came forward with a more straightforward approach and has won business away from Intel (see BW Online, 5/19/06, "AMD Inside,". Since being dropped by Dell (DELL) and IBM (IBM), HP is the only notable Itanium customer.
Analyst Linley Gwennap, head of The Linley Group, a Mountain View, Calif. consultancy, says HP could buy the Itanium business, and could indeed be the only logical buyer. "HP has made Itanium a major part of its strategy," he says. "I can't see anyone else buying a business that has only one major customer. It would only be a question of how much they would pay."