Non e' finita sui semiconduttori - gz
By: GZ on Venerdì 06 Febbraio 2004 18:21
(Nota: riportiamo questo tipo di materiale qui nel forum
perche' e' di supporto alle raccomandazioni che sono pubblicate nelle sezioni a pagamento. Ogni giorni ci riferiamo nel testo
delle varie sezioni come il Trading Track a questi post qui nel forum. Dato che per arrivare a raccomandazioni (si spera), utili
si utilizzano dei materiali di ricerca e dei report che si trovano
su altri siti e' utile ogni tanto, per chi fosse interessato ad
approfondire, riportarli anche qui )
A proposito del fatto che se uno puo'
leggere analisi serie fa meglio che puntare
soldi sulla base di un grafico sono tre anni che
ogni volta che trovo cosa Tom Kurlak
(il piu' noto analista di semiconduttori
a Wall Street, ritiratosi da alcuni anni) le riporto qui
e finora su quattro chiamate sul ciclo dei semiconduttori
ne ricordo solo una errata e nemmeno tanto.
L'ultima e' stata questa estate ed e' stata spettacolare
(i semiconduttori, lasciando da parte la povera Stm, sono
saliti in media del 50%)
Ora c'e' stata una correzione del 10-15% sul settore ed ecco
cosa dice Tom Kurlak. Due giorni fa l'ho letta e ho detto>
"... qui il Nasdaq rimbalza un poco..".
Kurlak non si basa sui grafici, ma su cose tipo i lead times di produzione, il ciclo delle scorte e simili stupidaggini economiche
che quelli che non comprendono dicono che tanto sono inutili...
Don't Bail on Semi-Equipment Stocks
By Thomas Kurlak
02/05/2004 10:32 AM EST
Basic commodity components are hard to come by.
Order lead times determine inventory levels.
A tightening DRAM market may be next.
I'm seeing a lot of tech analysts lowering opinions on semiconductor-equipment stocks lately in what is looking like a wholesale tech bailout. It seems that they're trying to avoid making the last-cycle mistake of staying too long in winning stocks.
But without any good reasons for getting out, they appear to be just jumping ship before something goes wrong. I understand the motivation, but I don't know why, at this early stage in the chip cycle, they would want to sell.
Yes, the Federal Reserve will be increasing interest rates before too long, but my experience has been that it takes a lot of rate hikes, at least four or five, to stop a bull market. In fact, in every bull market, investors learn to live with higher rates, eventually seeing it as another sign of the economy's strength. Inventory building because of stronger sales is usually what drives up the demand for money, and Fed rate hikes typically follow money demand up.
Speaking of inventories, I keep hearing about shortages for lots of low-tech electronic components, especially in Asia-Pacific markets. Makers of computers, cell phones, semiconductor equipment and even cars say they're having trouble getting quick delivery of basic commodity components, including capacitors, connectors, resistors, diodes and power transistors. China appears to be the new market sucking up the supply of these items. As I said last December, makers of these commodities look like good investments to me, and I have been buying capacitor maker AVX (AVX:NYSE) .
Also, I believe a tightening of the DRAM (memory) market is next, and, in fact, DRAM prices are firmer than most had expected for the post-Christmas period. Another company whose stock I own, Micron Technology (MU:NYSE) , says its DRAM prices are running about flat this quarter vs. last, which is a lot better than the 11% estimated decline predicted by the Wall Street consensus.
Tracking the Cycle
In terms of the semiconductor cycle and when it's time to sell, I've used my own decision matrix over the years, which comes from following lots of cycles. In late 2002, I outlined some of those key factors to knowing when to get into the stocks. Basically, they were indicators of when customers were buying fewer chips than they consumed and of when inventories were depleted.
For cycle tops, we need to track indicators of excessive demand. A good measure of that is order lead times. When lead times get out beyond 20 to 30 weeks, buyers are really just guessing what they need, so they usually overorder. But right now, lead times are only just starting to stretch out to 10 to 12 weeks; that's only for lower-tech commodity parts. Primary semiconductors, such as microprocessors, memories, analog and programmable logic, are still available off the shelf or within six to eight weeks.
Lead times are the best measure of inventory levels for semiconductors that I know of. As lead times increase, customers have to raise their chip inventory level to avoid out-of-stock situations that can shut down their assembly lines. At lead times of more than 20 to 30 weeks, you can be pretty sure that everyone is carrying too much inventory. Eventually, new capacity comes on, bringing down lead times and allowing customers to draw down their parts inventories. This is when semiconductor orders plunge and the downcycle begins. In my opinion, we are a long, long way from that point today.
Indeed, semiconductor-equipment companies are just beginning to see the order surge that precedes the next capacity expansion phase. It started last fall, first for assembly and test equipment, which is typical, and now it's picking up for front-end processing equipment. The leading U.S. test equipment company, Teradyne (TER:NYSE) , and the world's largest process-equipment maker, Applied Materials (AMAT:Nasdaq) , are the two stocks I own for the cycle.
In 2004, after some recovery in 2003, I expect to see overall equipment orders double and shipments jump 50%, both above consensus, due to the need for new capacity plus the transition to larger 300-millimeter wafers and copper interconnects needed for higher yields and better performance. In past cycles, it's taken at least two strong years of equipment buying for new capacity to outrun chip demand. That would get us to 2006.
Both Teradyne and Applied Materials are selling at fewer than 10 times potential peak per-share earnings this cycle of $3.00 and $2.50, respectively, if the cycle follows its historic pattern. For my money, it's way too early to start discounting that eventual peak from P/Es below 10, especially because the downturn was so severe that the upcycle could last longer than usual, taking peak earnings well above these estimated levels.