Mostly off Subject chat for Pat....
Regarding Small Caps & manipulating statistics to "prove one's point" (something I got rather good at many years ago in a survey research course)... as Mark Twain said: "There's Lies, there's Damned Lies, and there's Statistics". Always try to look at the raw data & the sampling methords.
Now as for RF Micro, it's interesting you would ask me about someone "in my own backyard". I admit I hadn't really given them much of a look, & I certainly claim no more than a passing acquaintance with gallium arsenide chip design/fabrication/markets, but I'll give you my thoughts - rudimentary as they are, and what facts I've rounded up. If and when I turn up more I'll let you know.
As you know their IPO was first week of June and they priced at $12 (higher than the price talk). It appears the overallotment was exercised by the underwriters, with Montgomery (soon to be Nation's Banc unit - folded into their Capital Markets Group) as lead, and Opp. in selling group. H & Q currently has a buy on them (I'll attach below) but I'm not real sure of the strength of H & Q's convictions here - I noticed they projected annual sales growth going forward 5 years at 30% annual - and the industry research I've looked at seems to project the overall GaAs industry will be growing at a 29.7% annual rate over that same 5 year span. There's those damned statistics again. Too close to be called an accident, it appears H & Q is projecting that RFMD will pace along with the pack. Admittedly though, a fast moving pack. It may certainly be assumed (and perhaps discounted because of that fact) that lead underwriter Montgomery concurs....
As you also know, RFMD traded yesterday at $23, a very respectable RamBus-like pop for one month. Still, "a rising tide..." and yada yada yada. The overall market HAS been very strong (particularly tech) over that period, and the IPOs that have recently been brought have been pretty much the pick of the litter, with several weaker co.s delayed or canceled. The investing public has visions of Ciena & Rambus and other sugarplum fairies still dancing before their eyes. This background undoubtedly has influenced the market action, whether fairly or unfairly I can't say yet... more on that later.
I'm about to expose my ignorance of the GaAs field.... As I recollect, there have been several "bulges" of speculative interest in gallium arsenide technologies over the past 2 decades. The fundimental advantage of GaAs being the ability to handle higher frequencies - thus, faster chips -, and the primary rap against it being that it is more difficult to fabricate and thus more expensive. In the early 80's the companies in this field were held up by military contracts, the military having great need for fast and more rad hard chips. The commercial communications market demand for GaAs was more embrionic. Expertise was and still is to a lesser degree concentrated in a few firms. This niche market status has tended to keep barriers to entry relatively high - one of the signs I look for in an investment opportunity. If the communications market - particularly wireless - is now expanding at a much more rapid clip worldwide (something I firmly believe it is) then these 30% and up growth rates should be sustainable for the overall industry; which brings us back to the company selection. Kopin currently has a very large share of the GaAs market, and Vitesse, among others. If I told you that I have enough insight now to be able to handicap among the contenders, I'd be lying (unless of course I could martial some convincing looking statistics!)
Foremost among the gaping holes in my knowledge here, I need to understand RFMD's current market opportunity better. I need to know their customers (haven't had a chance to read through the SEC filings yet, but I'll post them below). And what exactly is their advantage as "the only supplier of FaAs HBT components". They were a fabless designer of gallium arsenide hetrojunction bipolar transisters, and they used TRW's fab in Calif. for production. Now they want to build a fab in Greensboro, NC (note: that is in the "triad" [Greensboro, Winston Salem, High Point], not Research Triangle Park in the "triangle" [Raleigh, Durham, Chapel Hill], although it's only about 45 min. to an hour apart via I-40). Were there problems with the TRW arrangement or have they outgrown it's utility as they claim? I noticed that TRW is a minority investor in RFMD, and an outfit called ALN Telecom owns nearly 1 million shares. It appears that the new FaAs products are to be made in Greensgoro. The IPO raised $36 million and yet only $15 million is going towards the fab (the rest mostly paying down debt). Again my ignorance is showing but even in North Carolina $15 million doesn't seem like much of a fab. On a positive note, there was very little selling by insiders. Annual sales last year were reported at $30 million (not too bad if it were AMTX!).
To summarize a little, this could be a great outfit going forward (yes I do believe in the market opportunities for wireless local loop & PCS), and in hindsight I wish I was in on the public offer or earlier, but I'm generally skeptical of jumping in after a big runup like this unless I'm absolutely sold on the fact that the co. offers opportunity that cannot be duplicated, and I haven't seen that yet. If you see it, by all means let me know! And back to dreary statistics again, we all know where most IPOs traditionally finish one year after the offer.... Perhaps a better buying opportunity may present itself.
And now, here are some hopefully relevant links and attachments,
Regards, Buddy
techweb.com
edgar-online.com
techweb.cmp.com
September 09, 1996, Issue: 1023
Section: News
First Full Fab For RF Micro
By Jack Robertson
Greensboro, N.C. - The rapid growth of applications for wireless
components is prompting RF Micro Devices, a small, fabless chip
maker, to take the big plunge into manufacturing its own products.
RF Micro, which will post sales of about $30 million this year, said last week it will build its first full fab here to help meet escalating demand for its wireless devices based on gallium arsenide (GaAs) technology.
The company has used five foundries to build its GaAs and silicon
wireless components until now. Its GaAs HBT (heterojunction bipolar
transistor) devices have been made by TRW Defense and Space Systems
in Redondo Beach, Calif., which is transferring the production technology to RF Micro Devices for the new fab. TRW has also taken "a significant minority equity share" in RF Micro as part of the arrangement, said David Norbury, president and chief executive of RF Micro.
"The transaction arose because our need for GaAs HBT chips jumped so
dramatically that it exceeded TRW's capacity. We will continue to use TRW as a foundry, but we both decided that any expansion of capacity should be in a new fab operated by RF Micro Devices," Norbury said.
GaAs HBT chips are bipolar devices that have an advantage over
conventional GaAs MOSFET components because they are more
efficient and don't require a negative bias applied to gates, he said. Norbury said he believes RF Micro Devices is the first commercial producer of the HBT chips, although several Japanese GaAs vendors are said to be developing similar devices.
RF Micro Devices also has GaAs MOSFETs and some silicon units,
which will continue to be produced by four different foundries.
The defense and aerospace division of TRW has long been a foundry for the RF components company. The foundry deal between TRW and RF
three years ago was heralded as a successful move by a military
contractor to transition its defense system technology into the commercial market.
Copyright r 1996 CMP Media Inc.
You can reach this article directly:
techweb.com
March 17, 1997, Issue: 1049
Section: News
Kopin Responds To GaAs Pressure
By Tom McHale
Fremont, Calif. - Kopin Corp., a supplier of gallium arsenide (GaAs)
wafer-engineered materials, said it has tripled its capacity over the past year in anticipation of strong growth in GaAs-device sales driven by demand for wireless communications equipment.
The announcement is viewed by some analysts as a bellwether of
acceptance of GaAs technology by cell-phone handset makers and other
wireless vendors.
Kopin, Taunton, Mass., focuses on heterojunction bipolar transistor
(HBT) wafers - a segment of the GaAs wafer business.
"We provide over 90% of the HBT GaAs wafers purchased for use in
fiber-optic and wireless communications circuits," said Paul Mitchell, Kopin's chief operating officer. He said the company supplies such device manufacturers as RF Micro Devices Inc., Greensboro, N.C., and Rockwell Semiconductor Systems, Newport Beach, Calif.
In itself, the HBT GaAs market isn't huge. In 1996 it was worth $12
million, compared with a total GaAs wafer market value of $153 million, according to Earl Lum, an analyst who tracks the RF-device business for Dataquest Inc., San Jose.
But the growth potential for the upstart HBT GaAs process is big,
because it meets the requirements of two wireless technologies that were commercially deployed only last year: Code Division Multiple Access (CDMA) digital cellular networks and PCS.
Matt Micci, Kopin's vice president for sales, said that compared with the older, competing ion-implant MESFET process, HBT offers advantages that have a bearing on the CDMA and PCS technologies: stronger power gain, the ability to put out a stronger signal across a wider range of frequencies, and power efficiency sufficient to lengthen talk time.
Some observers contend that HBT is simply too expensive and too
immature a process to bet on today. "Cost drives everything in this
business, and HBT is more expensive than ion-implant MESFET, for
example," said Alan Morrison, an analyst at Strategies Unlimited,
Mountain View, Calif. "But the technology's advantages in stronger power gain and talk time may see OEMs willing to pay for it."
The decision to expand its wafer-engineered materials operation is a sign of Kopin's ability to commercialize promising technologies, according to John Fan, Kopin's president and chief executive.
"Over the past year, production orders have almost tripled, and with the purchase of new production equipment as well as operating our
production facility 24 hours a day, we have the capacity to keep pace with demand," Fan said.
Among the new equipment that Kopin acquired to boost capacity were
two AIX 2400 deposition systems from Aixtron Inc., Chicago. "That
represents an investment in the $4 million to $8 million range," Dataquest's Lum said. "That is a major commitment for a company that recorded sales of just over $20 million in 1996."
Lum said he remains to be convinced that HBT offers sufficient advantage to lead chip makers to switch from MESFET at this point.
However, he said there will be a market for HBT wafers from newer
GaAs vendors. "The old-line companies like TRW, say, probably
developed the technology out of their military business," Lum said. "But for newer commercial vendors it doesn't make economic sense to handle that part of the business."
Copyright 1997 CMP Media Inc.
You can reach this article directly:
techweb.com
And lastly, a little AMTX related news:
HAMBRECHT & QUIST LLC issued Company Report
Posted: July 7, 1997
On January 17, 1997 Hambrecht & Quist analyst Joe Noel issued a 6 page Company Report on ASCEND COMM INC. Report highlights: "The fiscal 1997 EPS forecasts were raised from $1.65 to $1.67, to reflect exceptionally strong fundamentals. As a result, the BUY rating was reaffirmed primarily because the shares appear to be trading at a discount. It appears that the company has landed U.S. West as a major purchaser of TNT in order to support its ADSL, Internet and Corporate Access strategies. " |
