Rairden, Eric, et al, I went to the stockholders meeting in SF yesterday, and took copious notes. I'll just roll them back out here, with not much editing or commenting. If anybody has any questions because I'm too terse, or something, please ask.
I'll summarize up front for those that don't want to shovel through the notes: very positive meeting: tough problems of the past couple years behind them, next 2 or 3 years look great for the industry and better for LSI, great product lineup from traditional LSI, and Symbios, Gresham on line, so let the good times roll! (Why was the stock down Friday?).
Dave Sanders, VP and General Counsel went through the mandatory voting stuff and then turned it over to Wilf Corrigan, who "MC'd" the whole meeting. He did call five of his staff up on stage after his opening remarks to field questions, where one of them might be more expert. Still, Wilf fielded most of them. Wilf's opening was a good 45 minutes, including some slides. I had never heard him speak in person. My first impression was that he's low key, almost too much so. But after a while, you get that he's one of the pioneers in the industry, a very much in tune and hands on manager, and a very good overall CEO, GM. (My wife said he'd be an excellent teacher, because of the way he can explain tech stuff to non tech people).
He started out talking about the 3 year industry slowdown we've just come through. He said it should have been over at end 97, but then Asia happened, which dragged it out for another whole year. He mentioned another thing, that the "Dell Model" got very popular last year, where customers wanted only 5 days or so of inventory on hand, so they doled out orders for smaller quantities in a JIT type model. That slowed product volumes for a while, but now the pipelines are full and should remain that way.
He then talked about highlights for '98, Q199:
- Acquisition of Symbios in Q3.
- Major restructuring in Q4.
- Gresham production started. LSI has 320 acres there, either 10% or 20% of the whole town's acreage. $4 billion spent to date on it. 330,000 square feet of plant space. There is one fab there now, space for as many as 3 more. Gresham was a drag on margins through Q199 because of underutilization. In Q2, Gresham starts pulling its weight, and margins should return to historical levels.
- A bit of chest thumping, about the stock: #1 gainer of any S&P500 stock (up 134%), through Q1! #4 gainer on the NYSE.
- In Q1, SEEQ was acquired. "This will enhance LSI's network space."
Then he talked about the future:
- Everyone is looking for 2 - 3 upcoming good years for semis. Predicted increases in revenue YOY have been talked in the 10 - 15% range. Now, according to Wilf, in backrooms, industry experts are talking more like 15 - 20%. And, he thinks LSI will do better than the overall industry. Inventories are low, there is broadbased demand. We may even see a capacity crunch. Wilf said he is even getting calls from CEOs of customers often, asking for priority on delivery. 1998 was about 1.5 billion in revenues. Goal of 2 billion in 1999, and Wilf says they're working on doing better than that.
Showed some slides.
- One about Moore's law, O.35 to 0.25 to 0.18 micron, # gates on a chip going from 1/2 million in '94 to 26 million in 1999, etc. (with 0.18). Boilerplate.
- Global manufacturing: Gresham will go to 0.18 (there now?). I think they might be going to reduce fab size in Japan because Gresham is on line. Don't quote me on this one. Gresham 8" wafers. Wilf, sounding like he was kidding but he was serious, said LSI would like to be about tenth to go to 12". Intel has said they don't want to be first either, so who is it going to be? IBM?
- Evolution of LSI chip strategy: was all gate array, all customer design, no LSI IP. Now Cell based, with some products part customer and part LSI design, some all LSI. Cell based is better (of course), and the IP on the part of LSI is valuable. He said Symbios is cell based in the storage controller chip designs, so they fit in perfectly in that regard. Question: will FPGA eat into LSI's cell based business? No, cell based ASIC is in the 60s % wise now, FPGE 20s, same in 2001 or so.
- Slide on what makes up a typical system on a chip (SOAC):
Start with a processor core, like MIPS or ARM
Add some IP logic, firewire or ATM
Add glue logic and achieve a playstation or router or server chip.
- Slide on 6 major markets, and customers:
1. Networking: Cisco, Nortel, Cabeltron, Intel, etc.
2. Storage, control chips (traditional LSI and now Symbios): Sun, Intel, etc. for control chips on motherboards.
3. Consumer: Set top boxes, Sony playstation, etc.
4. Storage systems: EMC, HP ,IBM, Amdahl, NCR, Walmart, etc. for storage modules or systems.
5. Telecom - wireless
6. Computer: Note that the prospectus has a page all about this slide.
- Slide on the internet. Use of it doubles every 100 days! $3 billion E-commerce now, going to $400 billion in 2002, several trillions in 2006. Wilf said everything LSI does is one way or another for the internet, especially Symbios.
- Symbios. Doubled in revenues in the last 18 months, expect to double again in the next 18. Symbios will be offering a 50 GB hard drive as a basic storage element (Seagate is their main drive vendor). They will put 10 of them, giving a half Terabyte of storage, in a package about 6 inches wide by 18 inches high by 10 inches deep. I saw one there, but not with 50 Gig drives yet. They had some other neat show and tell stuff, including products of their customers, like a Playstation and an SGI and Sun workstation. Symbios sells complete systems to some end user customers, like Walmart, who don't do any of their own OEM. I didn't know that Symbios sells to end users. Walmart tracks every sale that takes place, world wide, on Symbios storage. I think they use NCR servers. Walmart may be up to 100 Terabytes by end 1999. From my experience in storage capacities at large customers, this is huge! Symbios is HOT. I keep saying Symbios, it's now LSI Storage Systems.
I'm going to wind this down, too many notes. Some questions from the audience:
1. How about fabless? Is that the way to go? Wilf, 'fabless can turn into chipless in a capacity crunch.' Laughter. Another thing, with SOAC type chips, with digital, DSP and maybe mixed signal on the same chip you damn well better know what you're doing,or you really won't get any good ones for another reason: no yield. He's saying that the most complex chips really can't be sent out to a foundry because close interaction between development and manufacturing is required. LSI is very comfortable with their "fab-yes" structure. I think I coined a phrase, or have I seen it here? Fab-yes, hmmm.
2. Competition? In CPU areas, IBM. In telephone, telecom, networks, TI. Otherwise, Japan Inc., STM, C Cube. In storage systems, EMC, Clariion.
3. How about CDMA work? Wilf said LSI's CDMA chip, with DSP plus mixed signal plus the microprocessor core, is more integrated than Qualcomm's. OK, then let's get the stock up to a big fraction of QCOM! - my ed.
4. Playstation revenues going forward the same, less, what? Wilf - probably higher.
5. Copper chips from LSI? LSI has concentrated to date on getting lower K dielectrics for the G12 process, giving faster transistor speeds. They think this yields better performance than copper would to date. This is the same course taken by Intel so far - my ed. LSI will do copper on G13 or G14. Again, same strategy as Intel, if G13 or G14 are 0.13. because Intel is going to copper for 0.13.
6. Any Y2K effect on business? Maybe more business in H1 from businesses, particularly for storage. Y2K fears tend to make companies and people want more copies of everything. Whether on line or hardcopy, it results in more storage being bought. Also, customers may be ordering more chips out of fear of power outages, or whatever, that may take out their favorite chip vendor for a while. This fear buying may slow down in H2, but then LSI's consumer business traditionally picks up. Playstations, Set Top boxes, DVD, etc. business should keep it going.
Boy, that's enough!
Tony
|
