Mr. Luna Beam: What really stands out in your chart comparing LSI to Intel, Altera, Analog, Xilinx, Novellus , and Applied Materials is LSI's relatively low price to sales ratio in comparison to the other six companies. It is not unreasonable to suggest that LSI could grow revenues 20 percent in the next year and see an expansion of the price to sales multiple to 4.5 from its current 3.0. If this does occur, LSI will be priced in the mid 60s.
techstocks.com
Most interesting is the much higher multiples accorded the fabless PLD guys as opposed to LSI. While much of this can be attributed to higher anticipated revenue growth for Altera and Xilinx, I am not so sure that this will happen
Two thoughts: 1. It will be interesting to see how LSI's burgeoning ASSP business does in competition with and comparison to the PLD guys. This is something that clearly needs to be monitored. The next time I talk to IR, I will try and get a sense of the ASSP growth in the business ( I also expect Wilf and company to hold good to their promise to break out the revenues for the six divisions once a year) 2. If being fabless is such a great idea, why wouldn't LSI want to spin off the fab part of the business and keep the design part of the business. It certainly would provide for less cyclicality.
And that is the point about what makes LSI such an interesting company. In the good times, LSI can count on not only an enhanced bottom line coming from the fab part of its business. In the bad times (which include the just past double whammy of a poor semi market and the Gresham startup costs), LSI needs to worry about its fab costs. Thus, it would seem that LSI is perforce, one of the more volatile stocks in the semi business.
And this leads to Patrick's scenario of imminent shortage of capacity at the .25 and .18 level. IF we see a capacity shortage at these levels, what are the fabless companies going to do? While customers may wait for a superior (for the time being at least) product like VTSS gas chips, surely they will seek other means of filling their needs for PLD chips, and that will give LSI's ASSP products an advantage because presumably it will have less capacity constraint problems since it owns its own fabs. Moreover, at some level, you really can't be a player unless you own the fab--or at least it would seem.(Expect a sharp retort from brother Tang on this one.)
Two other thoughts: Why isn't LSI able to do foundry work at Gresham? I would think that over time, this might not be a bad idea--especially to buffer fixed costs during low periods of demand.
Also: What is the growth rate of MINT's revenues? Seems to me that this is the bridge that could lead LSI into a superior position to be an excellent provider of foundry work. Clearly the value accorded to design houses such as Cadence is much higher than three times sales.
What hold's LSI back. In a word (or two) gate array, gate array, gate array. With a slowing gate array accounting for less than twenty percent of the business (and some of that is in government contracts that aren't going anywhere), it is obvious that LSI's other business is going to grow significantly (as it has even during this past three year downturn). Thus, the market needs to break down LSI's components into valuations based up the following--1. High end chips (core business of standard cell and ASSP) 2. Storage (Symbios)3.design (Mint)4. Low end chips (gate array) and 5 foundry (potentially).
My feeling is that a fair valuation of these five factors gives LSI a price to sales range of 2.5 to 5 depending upon the type of semi market LSI is in. Thus, it would appear that even now after the runup, that LSI is nowhere near overvalued long-term. |
