SAN FRANCISCO (03/03/2000) - The evolution of Intel Corp. has in many ways paralleled the development of the technology business over the last three decades. In the '70s, it was essentially a memory company - its first product was a 64-bit static RAM chip. In the '80s, after its microprocessors were chosen as the brains for the first IBM personal computers, Intel went on to dominate the market for PC processors. In the '90s Intel moved beyond chips, producing more and more of a PC's guts: motherboards, adapters and other pieces.
Now that the Net has superceded the desktop for computing, Intel wants to build the heart of the new powerful networks just starting to come online. But at what cost?
That question lingered last week in the desert air above Palm Springs, where Intel held its annual Developer's Forum. More than 2,000 developers showed up to listen to technical speeches outlining Intel's new chips and products. In contrast to previous years, when virtually all the attendees were PC makers interested in the company's new microprocessors, roughly 25 percent of the developers at this year's event came from other markets. Many were networking product makers: the developers building next-generation networking gear for ISPs and wireless networks.
Following his keynote address, Intel VP Mark Christensen, head of the company's Network Communications Group, described some of the changes in the event over the past few years."This was originally a PC-centric event. About two years ago, we started to give people some more of the concepts we were experimenting with, like security issues and some more futuristic stuff. Now's when we start to roll it out in earnest."
Intel's drive into networking began in 1990 with the formation of the company's networking group. But in the past couple of years, Intel has created three additional divisions, all focused on networking. "In 1996, I convinced the executive staff that we could grow faster," says Christensen. "That accelerated in 1997 and culminated with a reorganization in the last six months."
The acceleration has become typical of the company as a whole. As one employee puts it, "If you get tired of working at Intel, all you have to do is wait a year, and it's like working for a whole different company." Intel quietly has been using its seemingly limitless resources - a $300 billion market cap plus $11.7 billion in cash - to acquire startups and launch new business lines. But now it must maintain its focus while rapidly expanding into networking, communications and services.
In a 1999 interview with The Standard, Executive VP Gerry Parker, the Intel exec charged with heading up the networking operation, described the company's mindset in the previous two years.
"The charter from Andy Grove and Craig Barrett was to start a new business, any new business," he said. "Just don't annoy our customers by selling branded PCs and don't get into anything too far afield, like the car rental business."
Intel still doesn't rent cars. But it is in the Web hosting business, putting it in competition with companies such as Exodus and Global Center. It also puts its brand on toys, like an Intel microscope produced in conjunction with Mattel.
Boldly tackling new markets is one thing. But the fact is that Intel's latest moves are fueled, in part, by fear. With its core microprocessor business, Intel dominates what is largely a commodity market - hungrier competitors are undercutting prices, and in the case of Advanced Micro Devices, even offering faster processors.
Intel also has been distracted by the development of its complex new chip architecture, known as the Itanium. The Itanium target launch, originally scheduled for early 1999, has slipped by a full year.
Even the flagship Pentium III, which is based on core features that date back to 1978, has been beaten up by faster and more powerful chips from AMD. A successor chip, code-named Willamette, should win back some of that market share. But Willamette is seen mostly as a stop-gap measure before the Itanium chip finally comes to market.
Wall Street has not been oblivious to the company's struggle to evolve. Mark Edelstone, a semiconductor analyst with Morgan Stanley, says that compared with other large-cap technology leaders such as Cisco, EMC, Microsoft and Sun, Wall Street has penalized Intel for poor execution. Currently trading at around $100, Intel shares have been as low as $50 in the last 12 months.
Although Intel doesn't break out financial results for its various divisions, the company's most recent SEC filings provide data for the new business groups - including the Wireless Communications and Computing Group, the Communications Products Group, the Network Communications Group and the New Business Group. In the most recent quarter, these groups lost $564 million on $1.2 billion in revenues.
The Network Communications Group has made 15 acquisitions in the last three years, according to Christensen. Another 10 are expected this year. "When will we be done with acquisitions?" Christensen asks rhetorically. "Is Cisco ever done?"
Other groups are even more aggressive. John Miner, head of the Communications Products Group, is trying to create a whole new product category. His group is developing products that sit between e-commerce servers and networking gear, working to speed up transactions. In existence for seven months, Miner's group is already shipping products. "I introduced the idea on July 5," Miner recalls.
"It was approved on July 11, and announced to the world on July 14."
Recent hires indicate that the company's networking drive has just begun. David Tennenhouse, who worked at the Defense Advanced Research Projects Agency and MIT on networking issues, recently joined the New Business Group, along with Dalibor Dado Vrsalovic, who came aboard after helping build AT&T's new network platform. A former AT&T colleague reports that Vrsalovic had grown impatient with the phone company's painfully slow rollout of new technology and preferred Intel's faster pace.
Steering the company through its latest metamorphosis is a group of key executives like Miner and Christensen who have been with the company for more than a decade. The absence of a rigid hierarchy also helps keep the company nimble. According to a profile by Tom Wolfe published in Forbes in 1997, Intel in its early days had no real management structure. While that is certainly not true today, the upper echelon remains streamlined.
Middle managers who at other companies would probably never speak to their CEO often report directly to CEO Craig Barrett, who has direct oversight of most of the new business groups. Each business division makes quarterly reports to top management to defend its performance. The upshot is remarkable freedom of action - and a corresponding burden to perform.
"Sure, we're taking some big risks," says Miner. "But it's still very safe for us to take risks. We still have the resources, and we have the people and the intellectual capital to do it. And of course, having a monster market cap doesn't hurt either."