Paul Engel. Re: BW Interview with Kumar. Here are some excerpts of an interview with Kumar on technology from BW online. Any thoughts here regarding Intel?
Q: Do we know yet when we will see a post-Y2K buying spree in the technology arena? Do we have to wait for Jan. 2, or could companies pull out their charge cards before then, confident that the situation is under control?
A: The consensus in talking to MIS [management information systems] managers is that they still can't quite get a handle on the impact of Y2K. Thus far this year, the consumer market has been the primary engine of growth for PC demand. Large corporate accounts have considerably moderated their IT [information technology] expenditures. The hope is that they will open their IT spending next year with the Windows 2000 upgrade. We should see that early in the year.
If not, one could make the argument that the traditional PC upgrade cycle, which has historically been about three years, has been extended. The ramifications of that would be fairly dramatic. Today, the corporate PC market is a three-year replacement market. That means that one-third of your entire stock is upgraded every year. If that gets extended to every four years, the total of the installed base coming up for upgrade will be only 25%. That could have major negative ramifications, but so far we are not seeing any indication of that.
One related point is the impact of the server market today. Most of the server deployment has been at the low end. These are one-way or two-way servers, meaning one or two processors. These are mainly file and print servers. From a management perspective, it has become a logistical nightmare having a multitude of servers running on different operating systems.
The trend today is to consolidate all these servers into higher-end multiprocessor machines with 18 processors and up. The primary beneficiaries of that are the Unix vendors such as IBM (IBM) and Sun Microsystems (SUNW). Today, IBM's mainframe-class machine System 390 offers five minutes of downtime a year. Out-of-the-box Microsoft NT systems have two weeks of downtime a year, so clearly they are not enterprise-capable. What we see happening is that the primary driver is E-commerce applications and other mission-critical systems, which translates to Unix, not Windows NT.
Windows NT will not be mission-capable for at least another two years, so Microsoft (MSFT) will miss out on this server upsizing opportunity. Today NT only supports two nodes, with each node reporting to a box. That allows two machines talking to each other. Windows 2000 will support only four nodes, which is not mission-capable. Just to give a comparison, the DEC system [now owned by Compaq] supports 96 nodes. Windows 2000 is not very scalable. The beneficiaries will be Sun and IBM on the hardware side and potentially Intel (INTC), if it can execute on the high-end servers. But on the operating system side, Microsoft will completely miss out on the upsizing opportunity.
Q: From an investor's point of view, are we at a good stage in the semiconductor cycle?
A: Most definitely. If you look at the history of the semiconductor market, it goes in a three-year cycle. Last year it was about $138 billion. The forecast for this year is that the market will grow to $144 billion. In a sense, we've had three years of zero growth. Today, the primary driver is sales of PCs. They represent 60% of consumption. But the growth going forward will be communications, both wired and wireless.
By comparison, the semiconductor content per PC, which today averages $500, will remain flat for the foreseeable future. Very few of the participants outside of Intel will show any meaningful growth going forward. The primary way to play the semiconductor market will be to focus on vendors that supply building blocks to the Internet. That includes Intel with its recent acquisitions. The others would be Vitesse Semiconductor (VTSS) and Broadcom Corp. (BRCM).
You also have emerging trends like reconfigureable computing. Vendors that would benefit from that would be Xilinx (XLNX) and Altera Corp. (ALTR). Last but not least would be mixed-signal analog renderers, such as Linear Technology (LLTC) and Maxtor Corp. (MXTR).
Q: We will soon be entering the PC holiday buying period. Have you seen any indications of how it will go?
A: If you look at it from an historical perspective, the second half represents 55% of annual computer shipments worldwide. Our research indicates that for the second half of this year, we will see a similar profile.
In terms of unit growth, we're seeing a healthy third quarter and a sequential uptick in the fourth quarter. That translates to 24% growth this year, to 113 million units, vs. 91 million last year. To put it in perspective, last year's unit growth was 13%. Next year, we are forecasting 17% growth, to a little more than 131 million units.
The flip side is that the value of shipments will grow only 12% this year, to $191 billion and to $204 billion in 2000. That means that the growth rate drops next year to about 7% because of price declines. Today the blended average selling price across all types of computers works out to $1,700. That includes desktops, portables, and servers. By 2003, we expect that blended price to drop to $1,400. The price declines will be most severe in the desktop market.
Copyright 1999 The McGraw-Hill Companies All rights reserved.
HL
|
