Intel over-valued; I humbly disagree. I apologize up-front for this being a rather long post -- but I heard this same argument from April of 1995 to December of 1995. I disagreed with that argument at that time -- and bought quite a bit of Intel for prices ranging from $45 to $55. I'm not a short-term trader -- so I'll be holding this for quite a while -- and thought it might be interesting for some folks to understand just why it is that I disagree with your posit that Intel is overvalued. I think I understand why you believe it is overvalued; however, I think that you aren't fundamentally taking into account the value associated with a monopoly on intellectual property. Put another way -- if Buffet bought technology stocks, he'd buy Intel.
There is a lot of macroeconomic information in your post that I don't want to get into (I'm not smart enough to predict this type of stuff -- one of the reasons I like Intel, in fact, is that they are primarily paid in U.S. dollars and I worry less about hedging on the dollars direction); instead, I'll just focus on Intel. Your comment that Intel is over-valued seems to be based on Intel's valuation versus other semi. companies and Intel's potential growth rates. Your comment that the semiconductor market is not recovering appears to me to be valid given what we've seen and I think we'll see on semi. BtB. What I would wonder is whether this is a good or bad thing for Intel -- because I don't believe that you can call out some type of monolithic semiconductor industry with consistent P/E and growth characteristics and not be wrong concerning the two extremes of that industry: companies like Micron and companies like Intel.
I'd posit to you that there are at least two semiconductor markets -- the "commodity" semiconductor market in which players have to compete on what are essentially fabrication capabilities and the semiconductor market in which the players have significant intellectual property. Certainly the semi. BtB, which consists of over 1/3 DRAM's (last I looked), is down -- and with the impact of new fab's coming on line will probably not get really healthy for some time. Given the massive price erosion seen in the commodity semi. market, we're seeing the PC players take prices down dramatically while still having superb earnings (e.g., Dell, Compaq). This lowering in pricing to some extent (and I'd be a trillionaire if I knew what it was quantitatively) increases the demand for PC's. Thus, it would seem to me that Intel wins from this phenomenon as long as it isn't seeing massive price erosion for its products -- which of course, it's not, because it has successfully branded itself and because the overall PC market has been (and seem to continue to be) moving upward toward higher performance. Bottom line here: pain in the DRAM (and other) market helps Intel, it doesn't hurt them (unless, of course, Intel stupidly buys a lot of DRAM on speculation -- but I think that they learned their lesson on this one).
So instead of looking at semi. BtB, I think you're better off looking at worldwide growth of the PC market. So far, I've seen no one who predicted this below 15% -- and the consensus seems to be 18%-19%. Fairly healthy growth -- particuarly when you consider the price erosion inherent in this number and then adjust your PC unit count upwards. And this 18%-19% is seen as increasing in future years -- and doesn't take into account microprocessor upgrade opportunities (which I think is going to really help Intel as time goes on and the installed base increases).
Now, given this, to me the biggest questions concerning the future of Intel come down to just a few:
1. Will Intel have a competitor capable of causing the
company to reduce its margins (or lose market share)?
2. Will consumers keep moving toward higher speed
microprocessors when they buy PC's?
3. How well will Intel handle the P7 transition?
Concerning #1, I don't see where Intel will have any competition that hurts them for the next few years. In my opinion, the current competitor that they have with the most potential to hurt them is currently Digital with its Alpha chip. Digital's strategy is to compete on performance. Of course, as long as most of the world is running Windows 3.1/Win95 and the like, that's not a huge concern since Digital can't run that software in a manner that people would want. If the world moves tomorrow to NT Workstation, this lowers the barrier for Digital (and others) to compete with Intel just on speed. However, with Intel's investment in branding coupled with their ability to invest so much in fabrication and R&D, I don't see this as a huge threat in this decade.
Concerning #2, I think this is a major concern for Intel. At some point the software out there reaches diminishing returns even as Moore's law let's you increase transistor counts, and indirectly performance, dramatically. This is where MMX is such a good idea. Of course, Microsoft's Talisman could hurt Intel here by knocking out that advantage -- but worse case Intel will be able to have "for free" (versus $300 or more for Talisman) some type of boost to keep customers upgrading and moving up the performance chain.
Concerning #3, I believe that while the P7 marks the first architectural change since the '286 (or further -- I just can't remember that far back!), from what I can determine Intel has an excellent migration strategy mapped out.
When would I sell Intel? If the story changes -- specifically, if condititions #1, #2, or #3 change.
Mark |
