As to buying all three (WIND, MWAR and INTS), that is exactly what H&Q recommended in their pre-Embedded Systems Conference spot report. I quote: "Certainly it is impossible at this time to predict which of the leaders might land the "killer" applications that overshadows the competition and launches it into a dominant position relative to the others. It is for these reasons that we believe that investors should take positions in each of the RTOS players Integrated Systems, Microware and Wind River Systems."
This also seems to be what institutions are doing, and why particularly WIND and INTS stock move similarly.
But while all three companies will probably continue to reward investors at above market rates, I personally have over-weighted WIND because I think it is in the driver’s seat. INTS is not a pure Embedded Systems play, so I loosely follow it but I do not own it. MWAR is pure, strong in Digital TV and wireless, and I think should be owned as a potential core stock.
Applied Microsystems is another Embedded Systems tool vendor recommended by H&Q, but I skipped over it when it IPO’d, and I don’t own it now for one simple reason. It does not provide an RTOS, and therefore will not benefit from future run-time license fees on unit production. Having an exponentially increasing stream of royalties is the exciting part of the ubiquitous computing story. Why buy an embedded systems company without it?
About Hitachi. When chip makers bring new processors to market they need to make sure it is compatible with mainstream development software. Otherwise it might be passed over for no technical fault of the processor, just because the development engineers insist on using a non-compatible RTOS and associated tools. On the other hand, RTOS vendors want to be on as many production processors as possible, so they do not limit choices of their customers, and so they can compete for new design wins in situations in which the hardware is pre-specified.
Actually it is more important to the chip maker than the mainstream RTOS vendor, because there are more of them and a lot more money gets involved in production hardware than development software and run-time license fees. This means that all chip makers have to contract with mainstream RTOS vendors to get them to port to each variety of their processor family, and this becomes an attractive source of revenue to mainstream RTOS vendors. As they become even more mainstream, they can negotiate ever more profitable porting contracts. And then afterward, the RTOS vendor expects to make design wins on the processor, depending largely on the chip maker to push the sale. Expect Hitachi to partner similarly with MWAR and INTS. Be alert to the possibility of yet another RTOS vendor announcing a Hitachi port.
If you think about this for awhile, you should realize how this relationship contributes greatly to the winner-take-all syndrome I wrote about before. Hitachi wants to contract with the smallest possible number of vendors to get an acceptable coverage of the potential market. Those they don’t contract with will be too non-mainstream to waste their own scarce resources to do the port gratis, so they won’t. Since these smaller companies cannot keep up the pace of maintaining availability on most processors, they focus on one or two processor families, and become less and less competitive. Without support for most processors, they will rarely even be considered for strategic relationships with large embedded systems application development companies. As they become less relevant, their chances of being sought out by chip makers to port to their latest processor will become even less. This spiral does not bode well for their continued survival.
The Hitachi announcement is important, but not as exclusive for WIND as you might think. However, taken together, Siemens, Hitachi, ARMS, Motorola, Intel, etc., and you can see why WIND is the market leader. INTS is close behind, and MWAR is rapidly increasing their ports.
About the only processor family that WIND virtually owns is the Siemens 16-bit family. They initially ported a custom RTOS to one of these processors under contract with GM (to be used in diesel injectors starting in model year 1998, and GM electronic transmissions starting in 2000, and possibly in all GM fuel injectors starting around 2000). The custom RTOS, scaled-down VxWorks, uses only 16 bits, and adds functionality useful when synchronizing rotary processes. This make the system potentially useful for many manufacturing process control situations and for other many mechanical device controllers. When the giant printing press company, Heidelberger in Germany decided to use Siemens processors, they no doubt knew of the partial VxWorks port, but they required a full port of VxWorks. Consequently, Siemens contracted with WIND, and only with WIND, for a full port of VxWorks. Nice to be so much in demand.
Keep an eye on the Embedded Systems Conference in San Jose.
Allen |
