DLJ's Nice Call on Intel
By Jim Seymour
Special to TheStreet.com
3/1/99 3:27 PM ET
Puh-leeze. Don't tell me you were surprised by Donaldson Lufkin & Jenrette's
downgrade of Intel (INTC:Nasdaq) to market perform from buy. There have
been signs everywhere of a slowing market for the kinds of chips Intel makes;
there have been signs of eroding (if still huge) market share for the company;
Intel's prices are clearly under pressure.
And you were surprised?
There were, of course, a lot of innocent bystanders -- tech-stock holders who
got whacked when a market leader got slammed. But, hey, that's the game.
And DLJ's downgrade -- to be followed, I suspect, by a few more over the
next couple of weeks -- has the "defensible advantage" of being right on the
button
Intel is one of those growth stocks I love, period. But not enough to own it, at
least not right now and probably not for the next couple of quarters and
maybe not for the rest of the year.
Consider six bad indicators for Intel:
Market growth slows. I've been saying for a year now that we'll see a
flattening demand curve for business PCs throughout 1999 as we approach
Jan. 3, 2000. Companies are spending their money on services, not PC
hardware; capital-investment budgets are being raided left and right to carry
the higher-than-forecast costs of becoming Y2K-compatible. Look at what's
happening with Dell (DELL:Nasdaq), Compaq (CPQ:NYSE) and
Hewlett-Packard (HWP:NYSE). This may yet be an OK year for them, but it's
pretty clearly not going to be the kind of bang-up year for corporate
multiple-unit sales they, and their shareholders, expected.
These PC manufacturers are Intel's most important customers by far -- to
whom else would they sell Pentium IIIs? -- and as go the Dells of the world, so
goes Intel.
AMD muscles in. For the first time ever, according to market researcher PC
Data, Advanced Micro Devices' (AMD:NYSE) sales in January in both the
business-priced PCs and the sub-$1,000 consumer PC retail markets
exceeded Intel's. In January, AMD topped Intel 44% to 40% in the overall
market, and in the burgeoning sub-$1,000 market, the tally was AMD 50%,
Intel 25%. With its K6-3 ("Sharptooth") chip now out (but not available during
January, and hence not represented in AMD's win), and the powerful
500-megahertz K7 coming soon and aimed squarely at the P-3, AMD has
finally become the serious threat to Intel it should been 15 years ago.
Yes, AMD's been the heartbreaker of all time for a lot of investors over that
time: AMD wudda, cudda, shudda -- but didn't. Now the numbers show it's
finally doing its job.
And, yes, the PC Data study covered only sales at retail, not direct sales of
large numbers of units direct to large corporate customers. But the K7,
positioned squarely against Intel's Pentium III, will be reasonably successful in
that field as well.
CPU prices falling. With that decline in demand, as market share erodes,
Intel's prices are slipping faster than usual on that deadliest declining curve --
driven down, of course, by constant pressure from AMD, whose new K6-3
400-MHz processor hit the market a week ago at just $284. The P-3 sells for
$580 to $825. 'Nuff said?
Worse, Intel has only two ways of combating market-share erosion: cutting
prices and spending more to increase its marketing support for all those
"Intel Inside" ads.
New product blahs. Intel has made a big marketing bet on the Pentium III,
which arrived last week to ... yawns and snores. Tests show this chip is only
marginally faster than the existing and much less expensive Pentium II and
AMD K6-3 CPUs. Intel's selling point for the P-3 is that it has 70 new
instructions, which are supposed to make both applications and Web sites
run much faster. Few existing commercial apps exploit the speed
advantages, and I can't see much difference in Web-site performance on the
P-3-equipped PC I've been using for the past two months. A
less-than-charitable observer might say Intel chose the Web connection to
promote the P-3 because the Web is hot, hot, hot and the P-3 is ... not.
Productus interruptus. The P-3 is just a transition chip in Intel's plan, and in
today's 400- to 500-MHz speeds, just barely makes that definition. (A
1,000-MHz version has been demoed by Intel execs, who said we should not
look for it anytime soon.) Intel's much-ballyhooed Merced processor, a 64-bit
CPU using much faster "IA-64" architecture, is the real play for Intel.
But don't look for the Merced to hit your desktop anytime soon, says Albert
Yu, senior VP and GM of Intel's Microprocessor Products Group. Maybe the
middle of 2000. In other words, late.
Intel's going to try to plug the obvious hole (read: market opportunity for AMD)
with the P-3/Xeon chip this summer. It's a great server and workstation chip
and will be better when tech software starts using Xeon-specific features. But
Intel will have trouble justifying its initially high price. And it's not the right chip
at the right time.
Fun with Andy and Craig on spring break. The Federal Trade Commission
suit against Intel starts a week from tomorrow in Washington. You can expect
Intel to be much better prepared than was Microsoft (MSFT:Nasdaq), which
never took its defense against the Justice Department's antitrust charges
seriously.
Unfortunately, the FTC has a much stronger case on the facts against Intel
than Justice was able to press against Microsoft. In fact, lead prosecutor
David Boies had to fall back on embarrassing Microsoft witnesses and
undermining their credibility as he built a case he hoped would stand up on
appeal. (Old Lawyer's Maxim, v. 1.1: "When the law is in your favor, argue the
law. When the facts are in your favor, argue the facts. When neither is in your
favor, stir up confusion.")
Intel is at real risk in this FTC action. The FTC's basic contention -- that Intel
sold microprocessors for 2% to 3% less to PC makers that did not form
partnerships with competitors such as AMD and National Semiconductor's
(NSM:NYSE) Cyrix unit; and that Intel cut off critical advance design
information to companies such as Intergraph (INGR:Nasdaq) as a means of
punishing them for relationships with other chip makers -- are a lot more
provable than Justice's mushy charges against Microsoft.
--------------------------------------------------------------------------------
The real issue for Intel is how to bridge the next 15 months or so till it starts
migrating us all into the 64-bit world. That bridge is going to involve some
painful and expensive charges for Intel -- and some painful and expensive
times for Intel holders. Nice call, DLJ. |
