(Briefing says don't buy semis, semi-equips, dd ever again - errors: grouping all semis as one, ignores impact of servers on INTC coming back, can always buy semi-equip - good businesses eventually etc. Usually institutional giving up on an entire group is a sign of an approaching bottom. LSI set anyway.)
---
Updated 02-Jun-98
Processor v. Bandwidth
There is a paradigm shift underway in technology from processor speed to bandwidth. You might say that this shift is already
clich‚, but you can also say that respecting it could have paid off handsomely in the market. Chips are out and
telecommunications services and equipment are in.
Factors Behind The Shift
When we talk about processor speed, Intel is obviously the first company that comes to mind, but this story does not end with
Intel. It extends to the entire semiconductor sector, the semi-equipment companies, and even has parallels in the disk drive
sector. Some of the themes that are relevant in all of these sectors are as follows.
Lack of software upgrades to spur demand for hardware upgrades.
Commoditization of products.
Oversupply.
Pricing pressure due to Asian crisis.
Demand weakness due to Asian crisis.
The Intel Story
Though the story doesn't end with Intel, Intel best exemplifies the problems afflicting all of these sectors. Their average selling
price is declining because the demand for processor speed is declining. This is not to say that unit demand for chips is falling; it
will in fact be increasing for years to come. But we have crossed a line where unit volume is not increasing enough to offset
declining prices.
In the old paradigm, consumers and businesses would be paying top dollar for the Pentium II 400 chip and Intel's revenues
and earnings would be soaring. In the new paradigm, demand growth is the strongest at the low end: PII 200-300, a segment
of the market for which Intel no longer has pricing power. AMD and NSM are now determining pricing in this market
segment. This shift is the difference between increasing revenues/earnings and decreasing revenues/earnings for Intel.
Forget Merced. The delay of this chip is not why Intel has fallen from over $100 last August to under $70 now. Merced will
still arrive and it will still generate substantial revenues, but it will not return us to the old paradigm. Yes, Intel is a great
company, but semiconductors are not a great business. Why own a great company in a bad sector when you can own a great
company in a great sector?
1,000 Words
Let's move beyond why you shouldn't buy Intel, or semiconductors, or semi-equipment, or disk drives, and talk about what is
worth owning. This is the flip side of our paradigm shift: who benefits from the bandwidth revolution? The obvious place to
look is telecommunications services and equipment. Find a great company in this sector and you will have a great company in
a great sector. The chart comparing the stock prices of Intel and Lucent since the beginning of 1997 highlights this point (both
stocks are set to an index of 100 in 1997). With Intel, you've had dead money; with Lucent, you've more than tripled your
investment.
The Telecom Story
Telecommunications is not the Answer with a capital A, largely because the latter two factors noted as problems for semi and
semi-equipment companies are also factors for many telecom companies, namely Asia (pricing pressure) and Asia (demand
weakness). Thirteen telecom equipment companies had earnings warnings in Q1, and six of these specifically mentioned Asia
as a factor.
Telecoms equipment and services are therefore the answer with a lower case a. This will be a winning sector in coming years,
but many companies are at substantial risk of missing estimates over the near term due to Asia. The best approach to this
sector is to draw up two lists, a near-term buy list for companies with insignificant Asian exposure, and a future buy list for
good companies that face substantial near term challenges related to Asia.
In addition to looking for some protection from Asia, we also look for a diverse product line. You can hit some home runs in
the telecom sector with niche players, but you will be a Dave Kingman-type hitter -- a lot more strike-outs than HRs. The
diversified players offer much less risk but still hold huge upside potential. We'll break them down into equipment and services.
First, equipment.
Lucent (LU): You saw the chart above -- this is the great company in the great industry. Great management, diverse
product line, and not much Asian exposure.
Cisco (CSCO): Call them a networker, call them a telecom equipment company, call them whatever you want, but
also call them a beneficiary of the bandwidth revolution, a company led by a CEO (John Chambers) who is one of the
best, and a company that will never fall behind the technology curve (what they don't have, they'll buy).
Tellabs (TLAB): Not as well known as LU and CSCO, but here's a company with a dominant position in the digital
cross-connect market and huge upside potential in the voice/data over Cable TV market and, best of all, no Asian
exposure.
Nortel (NT): We'd pick the other three first, but here is another diversified telecom equipment pick with little Asian
exposure.
Here are some ideas in the telecom services sector. Our criteria in this industry once again include limited Asian exposure, but
the other key is a focus on data communication over voice communication.
Qwest (QWST): With no investment in an existing (read: obsolete) network, Qwest is building a top of the line fiber
network from scratch and they have the best management in the industry, led by CEO Joseph Nacchio, formerly of
AT&T.
Worldcom (WCOM): Even without MCI's Internet division, WCOM will be a global leader in data communications
with its UUNet division.
Aside from these picks are some compelling companies that carry just a bit too much Asian risk for our tastes. When Asia
shows signs of turning the corner, these companies will be worth a closer look.
Advanced Fibre Communications (AFCI): Stock has been struggling a bit recently, and exposure in China will be a
significant drag if there is a yuan devaluation, but its UMC product has been dominant in the US market.
ADC Telecommunications (ADCT): Stock was pummelled after announcing it would miss Q1 earnings, but has
come rebounded a bit after beating its Q2 number. ADCT has significant international exposure, but is worth a look
once Asia settles.
The Right Side of the Paradigm
The whole tech sector has been trading poorly of late, and the correction has probably not played out yet, making all of the
stocks mentioned here risky near-term propositions. In another year or two, however, we expect that the graph of Intel and
Lucent for mid-1998 to 1999 will look very similar to the 1997 to mid-1998 graph, and that you could say the same of most
any graph comparing a semi or semi-equipment stock to a Lucent, Cisco, Tellabs, Qwest, or Worldcom. The paradigm shift is
a clich‚, but it's also true.
Parting Shots
It's been warnings season for a few weeks, but so far, those warning have been mostly small-cap players who badly
missed their quarter. Now, it's time for the big leagues. Intel, Motorola, and Compaq warned on March 4, 5, and 6.
The corresponding dates this quarter would be this Wednesday through Friday -- don't be surprised by one or two
big-name tech warnings. Intel already guided Q2 numbers lower with their Q1 report, but we wouldn't rule out an
additional warning given the lousy pricing environment for low-end chips.
Warnings scorecard: 11 tech companies warned in the first two weeks of the Q2 warnings season versus just four in the
comparable Q1 period. This week is already off to a bad start, with four tech companies warning on Monday alone
versus seven for the entire comparable week in Q1.
Direct subscribers to Briefing should be sure and check out the archives section of many Stock Service pages. We
have enhanced the usability of the In Play, Short Stories, Story Stocks, Stock Ticker, Tech Stocks, Market Internals,
and Earnings Surprise archives. The archives for Stock Brief, Trader's Edge, Earnings Calendar,
Upgrades/Downgrades, Splits, Economic Calendar, and Earnings Warnings will soon see similar improvements. Stay
tuned.
----
Shane
(Amazing how in a year everyone is now saying chip stocks are dead. I look at Briefing as a take on Institutional Interest. When they get bored/say never buy them again it is usually time to buy - but be selective.
I also have said before that I do agree with the overall idea of the paradigm shift and the commoditization of processor speed - its bandwidth stupid, not the processor.
INTC will eventually be back with chips for the server segment because even though they are toast with the low end (and therefore should receive a lower l/t market valuation) they have the opportunity if they don't mess it up of taking the Windows NT world by storm and that should help offset some of the low end business (in a sense therefore INTC also plays into the paradigm shift!).
Of course what Briefing fails to mention is that without a company like LSI or PMCS or VTSS or ALTR there would not be a CSCO or COMS or Telllabs or if LU did not have an ASIC house they would not be the LU of today. I've made the case before that the chip segment at 150 billion should no longer be considered homogeneous - this pavlovian market tends to group it as one because of the past history and correlation with PC demand - again over the next few quarters we shall just have to see how this plays out - what will be interesting is to see if there ever develops a disconnect between the PC and non-PC chip makers.
Anyway importantly for LSI all of this plays straight into this paradigm shift and will benefit - their networking segment doubled YOY + high-end storage is really an Internet play + communications with GSM & CDMA + new generation consumer electronics (STBs are really a bandwith play, DVDs and DCAMs are replaced products for old technology - related to bandwith - information explosion) etc etc - LSI's story will not be INTC's story - sorry Mr. Institutional Investor you are making the mistake of over-generalization - naughty naughty.
BTW I think the whole paradigm issue has to do fundamentally with the following: In the future you need a little highly-specific processing of a lot of information (the bandwidth paradigm) vs. in the past you needed a lot of processing of comparatively little information (the CPU paradigm).
Hence in the future there will be 80 ways to connect to the net - the PC will be just one.
) |
