Intel is Sizzling
by Rob Landley
Austin, TX (Aug. 18, 1998) -- I didn't have space yesterday to talk about Intel (Nasdaq: INTC), but it had some good news last week that I'd like to analyze here. And heck, the stock just keeps rising. Ever since Merrill Lynch analyst Tom Kurlak tabbed it for the low $60s, Intel stock has risen on a frozen-rope line from $70 to $90 today. Interesting.
Intel is currently working to fix a heat problem in its new 400+ mhz "Xeon" Pentium II chips, which occurs when the chips are used in multi-processor servers. And this is actually good news, Fools. A heat problem is good news? Why? I'm going to try to answer that today while recognizing that many of you may only be half-familiar with Intel's products and business. I'll do my best to keep it simple and clear. Over time, I think you'll be able to grasp its model.
Ok, why is the heat problem in the Xeon chip good news? Well, it's nice to see Intel concentrating on the high-end again through very powerful chips like the Xeon, and not just because that's where the profits lie. Intel's business model has always been to spend a lot of money developing new high-end chip designs and manufacturing facilities. It has matched this approach with selling its older chips at lower and lower prices as demand for them slackens. Once it can no longer turn a profit on the aging chips, Intel stops production and transitions to a more profitable use of that particular manufacturing facility.
Surprisingly, Intel diverged from this business model with its Celeron line of chips, which were designed from day one to be as cheap as possible to compete out of the gate with the low-end offerings from Advanced Micro Devices, Cyrix, WinChip, etc. The problem is that this subverts its profit model. How so?
Well, all microchips are fairly cheap to make. The substantial costs come in their design and in building out the production facilities. Once those are completed, the mass production of microchips is very inexpensive. This model is a little like that at Pfizer -- where they spend big money on research and development, but then end up making a fortune on the mass production of little pills (in Intel's case, little chips).
Now, generally speaking, the more advanced Intel's production facilities are, the lower its cost per chip is, regardless of the design. Here's where the problem with the Celeron line of chips comes in. Designing a chip specifically to be cheap to manufacture doesn't make a whole lot of sense to me. The cheapness of manufacturing isn't so important to Intel's business model as is the power of the design. So why would they design for it?
Here's one possible reason. Intel has chosen to compete on the low end, lower its enormous profit margins for a while, and undercut the competition on price. The only justification I can see for running the Celeron line of chips is to decimate its enemies. The other processor manufacturers, with razor-thin margins, don't have the money to seriously invest in new designs and new manufacturing facilities the way that Intel can. So, in the long run, I expect Intel to return to being Intel -- heading back into the higher-end chips -- after snapping at its competitors on the low end.
So, that's why I treat the news of heat problems with the Xeon as good news. It represents a careful focus still (and hopefully forever) on the high end.
But Intel isn't working in a vacuum. Lately, small companies like Advanced Micro Devices have been getting a lot of press. But in the microprocessor world, IBM is clearly the more substantial threat. IBM was the first to get a microprocessor design (a variant of its "Power PC" chip) to break 1000 MHz with conventional manufacturing techniques. IBM was the first to get a copper-based manufacturing process to work. And now IBM has invented a new "Silicon on Insulator" process to better insulate the circuitry in chips. Imagine all three of these techniques used together. IBM may be a dinosaur, but it apparently saw Jurassic Park.
Motorola has also been a strong traditional competitor of Intel and remains important. IBM and Motorola are both capable of competing with Intel in CPU manufacturing, which is Intel's real area of strength. Lately, the smaller chip manufacturers (AMD and Cyrix) have been cutting deals with the big two for help in getting their chip yields up. In this environment, I believe that Intel needs to keep introducing new high-end designs and manufacturing processes, and to let the trickle-down effect take care of the low end. Rather than backing up the plumbing with Celeron, I think they should just trickle faster.
Back to the Xeon. If you ask me, the heat problem is just teething trouble. Microchips, like light bulbs, get hot when they use electricity. They're designed with heat sinks and cooling fans; they're made to function at high temperatures; and they're made to work at the lowest voltage possible to get the job done. That's three ways Intel can address the problem, and they'll figure it out. They're concentrating on this.
The fact that the heat problem shows up in multi-processor systems is another good sign: Intel is pushing multi-processor systems. They are pushing power and speed. Like Gillette's new three-bladed razor, Intel is finding ways to get customers to buy more and more of its products. Computers can gain power by using more than one processor. The main limitation is changing the software -- especially the operating system -- to take advantage of the other processors. (More about that later this week.)
If Intel can push multi-processor systems hard enough at the high end, they should eventually trickle down to the low end. In any case, it means Intel sells more processors. And this is good. I think this, in part, is why Intel's stock has been on the climb during the past month -- and why Mr. Kurlak at Merrill Lynch may have to issue a mea culpa.
Below are some interesting articles on Intel for those interested. The stock was up about $2 today -- so celebrate by reading more! And Fool on!
--Rob Landley (Oak)
Other recent articles of interest about Intel:
Intel works to fix Xeon heat problem
New book on Intel's excellent management
Universities get funds for chip research
Scientists discuss impact of copper technology in microprocessors
AMD and Motorola team up
Intel's brand name isn't always enough
Stock Change Bid AXP +2 7/16 96.13 CHV + 13/16 80.94 CSCO +1 11/16 98.13 KO +2 1/4 79.06 GPS +1 5/8 64.56 EK +3 1/16 85.56 XON +1 3/16 69.56 GM --- 67.81 INTC +3 3/8 89.56 MSFT +3 1/16 107.31 PFE +3 7/16 104.63 TROW + 3/8 35.00
Day Month Year History C-K +2.32% 1.30% 17.07% 17.07% S&P: +1.62% -1.74% 9.98% 9.98% NASDAQ: +2.04% -0.92% 12.23% 12.23% Cash-King Stocks Rec'd # Security In At Now Change 2/3/98 24 Microsoft 78.27 111.25 42.14% 2/3/98 22 Pfizer 82.30 107.50 30.62% 5/1/98 37 Gap Inc. 51.09 65.13 27.47% 6/23/98 23 Cisco Syst 86.35 100.19 16.03% 2/27/98 27 Coca-Cola 69.11 79.06 14.41% 2/6/98 56 T. Rowe Pr 33.67 37.50 11.36% 2/13/98 22 Intel 84.67 91.44 7.99% 5/26/98 18 American E 104.07 100.38 -3.55% Foolish Four Stocks Rec'd # Security In At Value Change 3/12/98 20 Eastman Ko 63.15 86.88 37.57% 3/12/98 20 Exxon 64.34 70.19 9.10% 3/12/98 15 Chevron 83.34 80.50 -3.41% 3/12/98 17 General Mo 72.41 68.75 -5.05% Cash-King Stocks Rec'd # Security In At Value Change 2/3/98 24 Microsoft 1878.45 2670.00 $791.55 2/3/98 22 Pfizer 1810.58 2365.00 $554.42 5/1/98 37 Gap Inc. 1890.33 2409.63 $519.30 6/23/98 23 Cisco Syst 1985.95 2304.31 $318.36 2/27/98 27 Coca-Cola 1865.89 2134.69 $268.80 2/6/98 56 T. Rowe Pr 1885.70 2100.00 $214.30 2/13/98 22 Intel 1862.83 2011.63 $148.80 5/26/98 18 American E 1873.20 1806.75 -$66.45 Foolish Four Stocks Rec'd # Security In At Value Change 3/12/98 20 Eastman Ko 1262.95 1737.50 $474.55 3/12/98 20 Exxon 1286.70 1403.75 $117.05 3/12/98 15 Chevron 1250.14 1207.50 -$42.64 3/12/98 17 General Mo 1230.89 1168.75 -$62.14 CASH $94.76 TOTAL $23414.26 *Please note: On 8/4/98 $2,000 cash was added to the
portfolio for future investment. This will be reflected
in the numbers as soon as possible.
*The year for the S&P and Nasdaq will be as of 02/03/98