Shane: It's good to see you back. But I have some qualms about some of your post. Some of your “facts” aren't quite facts. And some of your opinions are also a little thin. For example:
1. “When anticipated cash flow wanes the stocks will get hammered as the risks of owning them increase. LSI raised cash in 1995 through an equity offering and so did many others - there was the cue to sell”.
Shane: If one had followed your advice, one would have recently sold LSI in the low 20s since LSI just successfully placed a major equity offering to reduce the debt incurred by the Symbios purchase. The market momentarily saw it as a sign of weakness as the stock dropped to 23, but LSI has rebounded nicely to its current levels. Clearly, this offering (the largest of its kind by LSI) was not a sign to sell. To the contrary, the offering reduced the interest rate paid on the debt—for a purchase that was an absolute masterstroke. I would add that I suspect that over time, LSI will sell the Wichita Storage Systems business at a significant profit over its purchase price, despite the constant blather about using the storage systems business as some sort of R&D lab.
2. “I think that LSI was capacity constrained in Q1.”
The primary capacity constraint for Q1 was due to a significant increase in the storage component business and a resulting shortage in the old Symbios fab. The vast majority of this problem has been ameliorated, and the production is moving to Gresham. Thus, in 2Q it is likely that despite the increase in demand, the lead times overall are actually becoming less and less, although there are clearly lead times that are increasing
3. “As I see it LSI has to produce 2.17 billion to be "just average". I base this on an old LSI base of 1.25 b and an old Symbios base of 0.6 b, split 75% semiconductor and 25% storage components (both for full year 1998) - now tack on 15% growth for old LSI, 15% for Symbios semiconductor, 30% growth for Symbios storage components, and 20 mil for 6 months of Seeq and we get: 1.25*1.15 + 0.45*1.15 + 0.15*1.30 + 0.020 = 2.17 billion.”
Shane: These figures are utterly irresponsible and demonstrate an utter lack of knowledge of LSI's basic business structure. Specifically, your entire premise is wrong because LSI Symbios semiconductor and LSI Symbios storage components are the same division. Thus, I sincerely question your ability to make any kind of reasonable projections on LSI since you lack even this basic knowledge. Obviously you cannot have two different percentages of growth for the same division. For the record, the two different divisions are Storage Components and Storage Systems (as well as Computing, Consumer, Telecommunications, and Networking). For the record, it's the Storage Components Division that is going gangbusters.
Furthermore, your beginning number of $.6 billion for Symbios is grossly misleading simply because a portion of the purchased technology was written off. Thus, any reasonable projection for Symbios would start with a much smaller base.
Moreover, your post that LSI “has to produce 2.17 billion to be “just average.” is outrageous. I have no idea where or what “just average” means, but it sure doesn't require a 26% growth rate in 4Q.or a year over year growth rate of 40% as you project. I have never ever heard any such predictions coming from anyone at LSI. Thus, I can only conclude that this is your usual wildly optimistic thinking.
It is not the first time you have pulled facts out of your hat. (For example, just after last Oct. CC, you insisted that the goal for R&D was 15% by telling me to check with Diana of LSI IR. It is and was 13%, just as I posted).
And going back two years, your projections have often been far too optimistic too soon on such issues as Fibre Channel, DCAM, DVD, and PSI as the following post #7977 of November 26, 1997 demonstrates.
“Still 390 big ones is nothing to sneeze at and in 1999, if DVD triples to 150,000,000, digital cameras double to 80,000,000 and PlayStation (hopefully) remains flat at 300,000,000 then LSI in 1999 will be generating a very fat 530,000,000 from these 3 chips alone.”
Message 2820750
This is not to say that these products don't have a great future but like many of us, you were far too premature in your assumptions of their contribution to the bottom line. In short, I think you just need to confine your optimism and clear brilliance to real facts and not figures that come from your own creation that are made of whole cloth.
Moreover, you harangue against MSFT and INTC may be a little late in the ballgame. Clearly, Linux is beginning to overtake MSFT in significant areas of the small server market, and as MSFT moves into the NT model for Windows 2000, it is clear that the days of the MSFT monopoly in the PC market are gradually becoming numbered. Linux is the child of the Internet, and Linux could well become the dominant OS within the next decade. After all, those upgrades that come from MSFT every other year have conditioned a large percentage of the buying public to upgrading their system. It would not come as a surprise that with huge storage capabilities being built in the average PC, that people begin to install two OSs on their computer. I mean you have a OS that is free versus a closely guarded secret. Can there be any doubt that over time companies such as Red Hat and Calebra will have the necessary suites to effectively compete with MSFT? Indeed, I would not be surprised to see MSFT offer some sort of hybrid Linux product. Similarly, it should be interesting as to whether Transmeta can put together a CPU that can read other types of CPUs. We shall see.
Where I do agree with you is the relative predictability of the PSII revenue stream—once we are assured of a certain degree of success for PSII. Certainly for the near future, PSI will provide significant revenues and will buffer any lack of significant revenue from PSII
Indeed, the singular attraction of LSI is that with the enhanced IP in their portfolio the shelf lives of their products will appreciably lengthen. PSI was the prototype and there could be a half dozen to a dozen to follow where LSI is clipping coupons for a three to seven year period. This bull run in the semis will actually mark the first time that true SOC will have a chance to show its stuff.
It is in this context that the fab partnership with Malaysia makes tremendous sense as LSI moves further away from the commodity end and more to the high-end Intellectual Property area. What I find unique about the Malaysia agreement is the fact that it is not a pure foundry agreement but a sharing of some IP. This allows for some buffering of the business during both the good times and bad times, since it allows less fab expenses during the bad times, but concomitantly allows Malaysia to peddle its own product line during the good times. The net result is that there should be more stability to LSI's earnings stream over an entire cyclical period.
As for CDMA, Wilf claimed at the annual meeting that they truly have a superior chip, and if that is the case, then they are off and running with last week's DSP purchase. I have to believe that LSI wouldn't have purchased the additional IP if they weren't buying a superior core. Only time will tell on that one. I would like to see the purchase price. To date, the scorecard is two wins (Symbios and Mint) and two don't-knows (SEEQ and ZSP).
As a sidebar I would note that in one of the responses to your post, I find it interesting that Patrick seems to think that LSI will be timing the use of 12” wafers at the perfect time in the next cycle. Reminds me a little bit of people who said the timing of Gresham was perfect, except that LSI clearly missed by a good three quarters, and when they did open, they immediately showed capacity constraints by ramping up too slowly. These things are so difficult to time, since you are timing both the supply side and the demand side of the market. Throw in the problems with ramping up a fab module, transitioning to 12", and product cycles that are relatively exogenous to overall demand, and this makes perfect timing extremely difficult.
|
