Indirectly related to ORCL's business. Barron's has an
interesting article about INTC. First it talks about it's
pc business. Then Barrett, ceo, talks about INTC's strategy
how to survive and future direction. In short, Barrett is
redirecting INTC into a new market, the Internet: banking
on the server market- Who will benefit most from it?
ORCL, of course, imo!
"Just take the rule of thumb that for every 10 PCs on the Internet you need one server," says Barrett. "I'm looking out and I see a billion Internet users. On that basis I'll need 100 million servers, roughly 20 times the number that exist today. It's a huge potential market." Industry analysts differ in their projections, but most agree that we are looking at a total chip server market which could quickly eclipse PCs in importance and size. "
In essence, Intel wants to move up the food chain and become a supplier of chips to servers and networking devices, the high-end computers that power the Internet. And given the sheer potential of the 'Net, combined with Intel's track record, Barrett's strategy looks smart. The server market alone is growing at around 35% a year, compared to just 15% growth for PCs. Yet this new strategy by definition remains untried. Powerhouses in the server business, like IBM and Sun Microsystems, are not likely to willingly cede market share to Intel. And heavyweight networking chipmakers like Motorola and Texas Instruments aren't likely to roll over and play dead, either.
Nor is it assured that the growth of the Internet will continue in hyperdrive. If the rollout of broadband access to the American home comes slower than expected, demand for Intel-powered servers and networking devices may not prove strong enough to make up for the slowdown in the company's revenue growth from the PC market.
"The simple fact is that while it may be dangerous to bet on Intel failing, its new ventures are anything but guaranteed," says Howard Anderson, president of the Yankee Group. "Investors think Intel is bulletproof, but it isn't."
In a very real sense, Intel has prospered over the last 30-odd years by following one basic tenet, Moore's Law. It was back in 1965 that young Gordon Moore, then an engineer at Fairchild Semiconductor and later Intel's chief executive, first postulated that technological advances would allow the number of transistors on a chip to double every 18-24 months, thus doubling potential computing power. Much as E=MC2 opened the door for nuclear power and the atomic bomb, Moore's Law provided a roadmap to success, and Intel executives took it straight to the bank. Beginning with the 8088 chip in 1981, Intel has churned out a steady stream of everfaster microprocessors -- 8086, 286, 386, 486 and the Pentiums I, II and III-each setting the standard for performance and each consequently selling at a premium price.
A basic pillar of this strategy, of course, was that PC makers, not to mention PC buyers, saw the faster chips as a "must-have" product. For throughout all the early PC years, there was hardly a time when software wasn't running ahead of hardware or when PC owners didn't find that the machine acquired just 12 or 18 months earlier could no longer run the latest versions of many popular programs. It was part of Intel's credo that everyone always wanted the newest and best chip. That is, until last year.
Exactly what happened in 1998 is something sociologists and economists will be writing about for years to come. Some say it had to do with the economic laws of diminishing returns and the near-saturation of the U.S. PC marketover 70% of homes with kids now have a machine. Others, like Intel's Barrett, cite the Asian economic crisis, which forced memory-chip and disc-drive makers to sell their products below cost just to keep their factories operating. Still others insist it all had to do with software, especially the fact that these "good enough" PCs happily run just about 99% of all the software on the market, including the browsers and graphics programs needed to access the Internet.
Whatever the case, cheap PCs became the standard overnight. Companies like Compaq and Gateway and scores of no-name brands began pumping out sub$1,000 PCs at a ferocious clip. And most were powered with cheaper non-Intel chips. As if that weren't bad enough, it soon became apparent that the cheaper machines were not so much attracting a new breed of skinflint customers -- as some in the business had hoped -- as cannibalizing the existing market for mid-priced higher-powered PCs. As a result, over the first half of 1998, demand for Intel's top-of-the-line Pentium chips faded faster than underwear at an orgy, knocking sales lower and cutting earnings by 36% in the first quarter and 28% in the second.
With its stock in a dive-Intel's shares fell to as low as 35 in mid-1998-the company's first reaction was to slash the prices of its older, low-end Pentium chips almost in half to make them competitive with chips from Advanced Micro Devices and Cyrix, which were being used to power these new cheap PCs. Intel's response was to rush to market with its new, relatively low-performance and relatively inexpensive Celeron chip. Though there were a few glitches along the way, that move, combined with further price cuts this year, subsequently won back much of the market share lost in those initial months. Certainly, from Intel's perspective, it would have been dangerous to surrender the low end of the market on a permanent basis, since that would have left AMD and others with a base from which to expand up into Intel's main market.
Intel maintains that the margins on its low-end Celeron chip are not that different from those of its high-end chips. Perhaps so, but that still means that the company has to sell five $100 Celeron chips it makes to one Pentium III. And with sub-$1,000 PCs here to stay and the market growing overall at just 15% a year, that limits the long-term potential.
Barrett, a 25-year Intel veteran who was named CEO in May 1998, first made his mark at the company in the late 1980s, when U.S. chipmakers were under attack from their Japanese competitors. At the time, Intel was so battered that it was forced to lay off 20% of its work force and shut down its memory-chip operations just to survive. Rather than focus on what rivals like Hitachi, NEC and Toshiba were doing wrong-many in the industry accused them of dumping products below cost-Barrett, a soft-spoken former Stanford professor, took a close look at what they were doing right. Dumping or no, he recognized that his Asian competitors were turning out silicon cheaper, faster and better than anyone else.
His response was to go on a whirlwind tour, visiting U.S. chipmaking equipment companies to get detailed descriptions of just how Japanese chip plants used their machines, touring plants owned by Intel's Japanese allies, and even going out of his way to get reports from chip customers of what they had seen while visiting Japanese facilities. His conclusion was that while Intel still excelled at chip design, its manufacturing process was downright inefficient. It was Barrett who led the effort to overhaul manufacturing from the bottom up, setting the stage for the company to become widely recognized as the single most efficient chip maker of the 'Nineties.
"For the previous 15 years or so, no one had really paid attention to the idea that we needed to be efficient," says Barrett. "It was enough to be a technology-creation machine."
The challege Barrett faces today is less specific than the one he faced a decade ago, but it is no less urgent. "If life at Intel was based on just the sub-$1,000 PC market, it would be painful," he concedes with a grin. He adds, however, that there are new software applications out there on the consumer horizon, like speech recognition, digital imaging and video conferencing, that will require high powered PCs. And as they take hold, he says, PC buyers may once again start moving up to higher-priced models.
That said, he's not waiting around for the high-end PC business to come screaming back. Intel's next growth engine, he says, will be the Internet. "From our perspective, the continuing boom in Internet traffic will create an immense market for an infrastructure of workstations, servers and networking devices. Job 1 at Intel is to be the building-block supplier of the Internet economy."
Consider the server market. These high-end computers form the backbone of the World Wide Web, not only storing and transmitting the data, but also controlling, policing and supervising the flow of information-plus operating millions of Web pages. Barrett, who can be excused for holding an optimistic view of the market potential, figures that the number of servers needed to manage the 'Net could grow exponentially over the next five years.
"Just take the rule of thumb that for every 10 PCs on the Internet you need one server," says Barrett. "I'm looking out and I see a billion Internet users. On that basis I'll need 100 million servers, roughly 20 times the number that exist today. It's a huge potential market." Industry analysts differ in their projections, but most agree that we are looking at a total chip server market which could quickly eclipse PCs in importance and size.
|
