Most of the six elder statesmen of the
have reported earnings in the past few weeks, and the news was anything but gray for the three largest.
all reported impressive gains in revenue and earnings. Only
reported a loss -- not to mention its 12th consecutive quarterly decline in revenue.
Next to report are
on May 11 and
on May 13. Early indications suggest more of the same: The recovery in information technology spending is fully under way, and while stock prices already reflect much of that good news, both fundamental and investor momentum should remain on track.
Government data suggest that inventories are being rebuilt and that demand from consumers and businesses is modestly advancing. A sustained reversal in job growth, even if accompanied by higher interest rates, should accelerate both trends. Most personal computers are bought for new employees, and for better or worse, the tech economy still depends on the sales of those old workhorses.
Hardware Takes the Lead
What have investors learned?
- Over the past 12 months, hardware has become the star of the show. Intel shares have gained 45%, Sun has risen 29% from an extremely low base, and Dell has advanced 23%. Trailing badly are the software guys, with Microsoft shares advancing just 7% in the past 12 months, and once-mighty Oracle up just 2%.
- Since April 2001, hardware also holds the lead, with Dell up 38% and Intel down 9%. Meanwhile, Microsoft is down 18% and Oracle is down 28%. Sun has become inconsequential over the time span, having lost enterprise software business to inexpensive Linux providers and enterprise semiconductor business to low-price leader Intel. Sun shares are down 75%.
- Since April 1999, however, Oracle shares are in the lead with an 82% gain, followed by Intel with a 10% loss, Dell with a 13% loss and Microsoft with a 32% loss. Meanwhile, Amazon.com (AMZN) - Get Report and Yahoo! (YHOO) fared even worse, with declines of 45% and 35%, respectively.
This was not the picture that many expected to emerge in the late 1990s, when experts considered software the apotheosis of technology. Hardware was considered lifeless metal that only became relevant when sparked by operating systems and applications. Internet software was the "the dot in dot-com," as Sun put it.
Why did large-cap hardware stocks charge to a surprising lead in the past few years? The answer is fivefold. Hardware got smarter, as engineers began to design more software into their chips and boxes. Hardware makers did a better job fending off trustbusters in Washington. Hardware makers started the period more undervalued. Hardware engineers proved better at squeezing costs out of their businesses. And perhaps most important, hardware makers proved the better marketers. The hottest new technology of the past few years, wireless broadband, was brilliantly sold to the public by Intel as a function of its chips. Think of it this way: When you bought a Dell laptop with a Centrino wireless chip to surf the Web at Starbucks, you probably couldn't name a single piece of software that made the computer or Wi-Fi hotspot work.
A United Future
Now an innovative initiative gaining traction may help hardware and software makers advance to the next phase in tandem, as they will depend on each other more than ever. This new, new thing is "grid computing," which is a variant of an old, old thing once called "utility" or "distributed" computing.
Grid computing has many definitions, but it's easiest to think of it as many small computers acting as one large computer. These computers can be in one room at a company or university, or spread out over a region or country. It's sold as a lower-cost way of doing large-scale computing, and that's what it is. But grid computing's primary benefit is its ability to provide computing power on demand for varying intensities of a user's needs.
When you switch on a light in your office, for instance, you get only as much power as you need for the light bulb in that space. You don't need an entire power plant in your building to supply that juice, as you depend on the electricity grid of your city to provide juice as necessary, either from local power plants or ones many hundreds of miles distant.
But when you turn on your Pentium 4 computer to access your email, you have the technology equivalent of a power plant on your desk. You only need a fraction of your computer's capacity to do email, and the rest of its ability, referred to as "cycles," are wasted.
Grid computing provides ways for businesses to harness those wasted cycles that are sitting around on employees' desks. After everyone has gone home, for instance, grid-animating software from companies such as Oracle can put all of those idled computers to work as if they were a single, extra-large parallel-processing computer -- boosting "cycle utilization" from the low teens up to 80%.
Grid computing also provides a way for companies to buy a roomful of relatively inexpensive Dell "blade" servers for mission-critical computational functions, instead of a single expensive IBM mainframe. One example would be determining the most efficient way to move a container ship of sneakers from the Seattle docks to 175 retailers in the Midwest. After the shipping conundrum is solved, the Dell blades can be divided up into smaller clusters on the fly to solve six different smaller tasks that could not be run on a mainframe.
While the potential for grid computing has been discussed for years, it's only recently that hardware such as low-cost Intel Itanium-based servers, storage-area networks and Oracle software has made them feasible. So-called distributed computer schemes have long allowed computer clusters (running your accounts receivable and sales reporting software separately) to basically pass messages back and forth with each other. In contrast, Oracle Vice President Robert Shimp said his company's new software now enables Information technology (IT) managers to manipulate those clusters as if they were one big computer.
Because of the complexity of making various vendors' equipment work together, technology companies this year formed a consortium called the Enterprise Grid Alliance to develop tools and standards for developers and users. Oracle announced last week that it joined the group; other participants are
, Cisco and Sun.
Companies that should be winners in a new golden age of grids would be low-cost "blade" server makers such as Dell and Hewlett-Packard; software makers Oracle and
; storage-network equipment makers such as
; data-warehousing service providers such as
; and many private companies working on various parts of the solution. These include
, which may go public in the coming couple of years.
Mark Anderson, publisher of
Strategic News Service
and a veteran observer of technology trends, said he believes grid computing has evolved out of hype mode into a legitimate business weapon for large companies trying to recoup the investment in their existing hardware. He noted that the first interuniversity grid is currently being built in the Research Triangle in North Carolina at a cost of $30 million.
In short, grid computing could finally deliver much of the cost savings and networking power that Internet visionaries first promised enterprises many years ago: A transformative technology that makes both individuals and companies more powerful together than they would be separately.
For at least the rest of this year, I'll watch for new developments. If you have any comments on how grids are being developed at your company or school, email me at
Jon D. Markman is publisher of
StockTactics Advisor, an independent weekly investment newsletter, as well as senior strategist and portfolio manager at Pinnacle Investment Advisors. While he cannot provide personalized investment advice or recommendations, he welcomes column critiques and comments at
firstname.lastname@example.org. At the time of publication, Markman owned or controlled positions in the following securities mentioned in this column: Microsoft, Starbucks, Dell and Cisco Systems.