Thanks for your assessment of Radisys, x86 and the Microsoft threat. Microsoft gets lots of attention on this thread, and probably always will, because MSFT does represent a scary threat; albeit so much whenever we stop and look at it rationally. Nevertheless, we all fear the day Bill Gates wakes up and discovers embedded systems like he discovered the Internet in 1995.
In my mind, one of the points you made about x86 embedded systems deserves emphasis. You wrote:
>If our quantities were higher, and we were going to build our own,
>the x86 would not have rear its head.
Quick and dirty embedded systems can splurge on real-estate, OS license fees, and squeeze software development time. But should production volume increase noticeably (or if it is anticipated to be large to begin with), costs and form factor become important. This often requires that the design must be value-engineered to reduce cost and meet a myriad of other requirements. Just the licensing cost alone of embedding DOS or WINDOWS NT or even CE mitigates against MSFT in high-volume, embedded applications.
For example, in one of the articles Mark referenced to you, it was pointed out that Pentium II cartridges are poorly designed for embedded applications. Intel promises next year to provide more accommodative packaging for the embedded marketplace. Are these processors running a MSFT OS a threat to embedded RTOS vendors? Not as long as Pentium II processors cost in the hundreds of dollars? Not as long as MSFT demands high licensing fees. And certainly not as long as a myriad of other functional considerations must be met for high-production designs, only one of which is hard real-time.
A final point on Radisys and competition. One of the biggest traps for the high-tech investor is to buy a company which enjoys an apparent monopoly, but actually simply hasn't yet been faced with much competition. In fact, this is probably the single, biggest trap for the unwary. The seasoned investor always asks why a company gets to have high profits, which implies high gross margins. The answer must be that there are considerable barriers to entry of competitors, or else the investor will discount the profits, and maybe the stock. Without barriers to entry, the investor knows full well that competition will soon emerge, forcing the company to squeeze margins. (I probably spend more time thinking about this one consideration than anything else when evaluating new stocks and stocks I already own, including WIND.)
One of the best examples of this phenomenon is EXBT. When EXBT introduced the 2 gigabyte 8mm data cartridge, they did well and the stock did well. Buy an 8mm by any brand name, and underneath was an EXBT tape transport engine (made originally by Sony). Unfortunately for EXBT, high-density tape systems seemingly grew out of the woodwork, because EXBT had no way to stop the competition from entering the business and taking share. Years ago EXBT stock tanked, and has continued to disappoint ever since, even as the company continues to upgrade their products and increase revenues.
I was corrected in an earlier post when I said only monopolies can beat WIND's 18% net after tax profit. Its true, some non-monopolies can also, but as a general rule, not for long. It is generally true (Economists must have a law about this) that the market will not allow any company to benefit for long from excessive margins if it makes economic sense to compete for the business. This suggests that if Radisys continues to grow its business profitably, and there is little to prevent the emergence of competition, then you can expect Radisys to have its margins squeezed by competition. Without barriers to entry, Radisys becomes a questionable buy.
In contrast, some of WINDs greatest attractions are the inherent barriers to entry by competitors in its primary markets. Imagine yourself knocking on NASA's door, and saying, "I would like to show you my new RTOS and IDE. I will sell it to you for half what you are used to paying. All you have to do is spend hundreds of thousands of dollars to verify that my RTOS is up to your standards, and additional hundreds of thousands of dollars to re-train your development engineers to use my tools. The fact that all your engineers are religious about VxWorks isn't a problem, is it?" Even the government rarely falls for that line anymore?
Allen |
