That's the lesson from the recent announcements out of Intel that itits two unprofitable communications groups and to cover losses associated with an acquisition.
Even though it is the largest, richest and most successful chipmaker on the planet, Intel has struggled to become a titan in the market for communications chips.
In a way, a company merging into a new market is like a person trying to change personality.
Many of these acquisitions have since been folded or spun out at a loss, according to sources. The newly merged units--the Intel Communication Group, which makes silicon for routers and other telecommunications equipment, and the Intel Wireless Communications and Computing Group, which makes cell phone chips--have together amassed $2.6 billion in operating losses since 2001. Several management changes have occurred.
"Becoming the world's leader in processors for the PC was part of their destiny," said Steve Tobak, a principal consultant at Invisor Consulting. "This is different. There is the question: whether their heart is truly in it...Networking is not in their blood."
Interestingly, Tobak's comment isn't a recent remark. It dates back to 1999, when Intel's buying bender had.
So, what went wrong? The downturn in the communications market certainly hurt. Intel's two communications groups swung from a $927 million operating profit in 2000 to a $991 million loss in 2001. (Those numbers, however, also reflect results from Intel's flash memory group, which has been around for years and is said to be profitable.)
The push, though, was hurt by the usual factors. Entering into new markets, especially through a series of acquisitions, can be an organizational nightmare.
An early sign that Intel was having a little trouble in this department was the disappearance from the company of Robert Pepper, CEO of Level One Communications, a maker of networking chips
The truth turned out to be much simpler. Pepper had retired six months earlier for genuine personal reasons.
More importantly, newcomers lack the tribal knowledge--the subtle and often unarticulated dos and don'ts of a particular field. This is particularly a problem in communications chips.
"Radio frequency design is magic. (Communications chip designers) are pretty esoteric people as well," Stephen Pawlowski, an Intel fellow, said during a roundtable at the recent. Michael Hogan, CEO of Sirific Wireless, nodded in agreement.
In a way, a company merging into a new market is like a person trying to change personality. We all knew someone who tried this in high school: One week, a Trekkie; the next week, wafting to algebra class in a black cape and Goth makeup. A roommate back in grad school bought a bunch of video tapes of self-improvement guru Tony Robbins, but he never became personally empowered enough to remember to pick towels up off the bathroom floor.
AT&Ttwo PC companies in its attempted conquest of the desktop market, which ended in embarrassment.
Some metamorphoses do work. Back in the '80s, Intel got out of the memory market to concentrate on PC processors. The results were astounding: The Japanese memory makers found themselves locked into a profitless price war, and Intel emerged as one of the most powerful companies in the information technology world. Intel's then-CEO,, went onto become a business legend.
Grove's dramatic shift was brilliant, but circumstances helped quite a bit. The PC market was just taking off, and only two companies--Intel and Advanced Micro Devices--had a license to manufacture Microsoft-compatible processors. At the time, AMD was floundering. It could not have taken over the market, and Intel would inevitably have dedicated more energy to processors. The shift would have happened, anyway.
Intel's push into communications is not over. Ultimately, pricing triumphs in semiconductors, and Intel can produce chips for less than most of its competitors--which should attract customers over time. The company has also mellowed a bit. Way back when, its sales reps were known as "FIGs," for @#$*! Intel Guys. Now, Intel offers to build servers or other equipment for manufacturers short on design budgets.
Nonetheless, Intel's colonization of communications is going to take a long time.