I'm back for a couple weeks, then off again - and I was able to hear the CC. Everything you said was correct, but allow me to elaborate on most items:
1. Re the auto market: In general, Ron said that they've found that the auto market for embedded products moves very slowly, and the direction for the industry is being established in Europe. Most importantly, he said that WIND made an investment in a German company (can't remember the name) that has allowed WIND to make recent inroads into the auto market.
WIND has been active in the auto business since GM announced they selected WIND for diesel injector systems in 1994, to begin production with model year 1998, which I believe has slipped to model year 1999. The problem has always been auto companys' reluctance to pay for IP (i.e. royalties). Ron said WindStream, the customized RTOS for GM, was selected for electronic transmission, but there have been no other auto announcements for under-the-hood process controls. The deal, including an investment, with the German company 3Soft, dreisoft.de , involves 3Soft porting VxWorks X WindStream to chips coveted by automobile companies for under-the-hood control, and should help secure numerous auto design wins. Apparently 3Soft has been a leader in helping the industry develop the auto operating system open standards called OSEK. (This has been a story developing mainly in Europe slowly over many years, but apparently is coming to fruition.) Notice that 3Soft is primarily OS agnostic, willing to work with Windows CE , VxWorks, Linux, etc. Don't forget that even Microsoft would shudder at the thought that Windows CE might be used for an under-the-hood control function.
The thing I found most intriguing (and maybe I misunderstood, so someone please correct me if I'm mistaken) on the call was Ron's statement that I2O has been held back because Microsoft hadn't released an OSM (?), and that I2O would now begin to ramp up in earnest because MSFT finally released the OSM recently.
Clearly Windows NT 4.0 has been patched with a suitable OSM, or at least a means of tricking the OS into working, perhaps reluctantly, with I2O. Otherwise, how have all those transaction speed records been set, and how have all those vendors been able to demonstrate the benefits of I2O? Further, Microsoft sits on the I2OSIG steering committee, and has endorsed I2O by agreeing to include the I2O OSM in Windows NT 5.0 when it debuts next year. SCO Unix and Netware are either in production or soon to be in production with I2O extensions. So what was Ron talking about?
The fact is that Microsoft has not been as enthusiastic about I2O as Intel. If Microsoft wanted I2O half as much as Intel, then the company would have produced an OSM patch immediately for Windows NT 4.0. Frankly, they have been disappointedly slow about joining the party. This reluctance in turn has slowed the production ramp up of I2O, as the industry waits to see how the giant accepts I2O.
But the fact is that Microsoft needs I2O as much as Intel, in order to build mission critical server systems for corporate America, and Microsoft's business model, like Intel's, is hugely dependent on deeply penetrating the server market. As a consequence, Microsoft has overcome its reluctance and issued a pre-production release of an OSM patch for Windows NT 4.0, according to Ron, with the production release scheduled for late June or July. This means all the add-on device companies will begin ramping up production on I2O-centric devices, knowing that the standard has been blessed by the leader. This is why Intel now indicates that I2O production volume will begin to ramp up in this quarter, ending June.
Ron also said that I2O revenue would begin to show up in Q2 and Q3, and that the revenue would be "material" which he later defined as "in the millions."
Actually, they will show up in Q3 and Q4, although they may show up in Q2 if Intel were to report production volumes sooner than the 60-day requirement. Worst case is two production quarters this fiscal year, which nevertheless is presumed by Ron to be material.
Sheila Ennis, the H&Q analyst, was pushing Ron about I2O because moments before the CC she issued a spot report in which she said, "Though the company outperformed our estimates by $3 million on the topline and 2 pennies on the bottom, we have added only $1 million of revenue and a penny of EPS to our fiscal 99 forecast. Due to continued uncertainty as to when I2O will begin to contribute in a meaningful way, we are lowering our fourth quarter revenue and earnings estimates."
In effect, Sheila substituted increased overall performance for disappointing I2O revenues for FY1999 - a wash. Now she knows that I2O will be material, and should be re-included, upping her estimates even more for FY 1999.
In addition, despite Sheila's misstatements about CORBA's importance and INTS' pRISM+ long-term favorable position relative to WIND's Tornado, and concerns about I2O timing, the 3-5 year revenue trendline she forecasts has been quietly shifting more and more in favor of WIND. For example, as late as January of this year H&Q continued to forecast 3-5 year revenue growth for WIND, INTS and MWAR as 30%, 25% and 30%, respectively. The forecast for WIND was upped to 41% in March. In April, following the ESC in Chicago, these forecasts were changed to 40%, 20% and 25%, where they stand today.
Since H&Q sees the embedded systems sectors growing at a 30% CAGR, this means that H&Q realizes that both INTS and MWAR will continue indefinitely to loose market share to WIND, with their hopes tied solely to a growing sector. H&Q sees WIND as growing market share and revenues around twice the rate of INTS, while also obviously doing well in the face of the Microsoft threat.
Its time for Sheila to join the crowd and issue a Strong Buy for WIND.
He also said that all the hoopla over embedded Java doesn't mean a whole lot right now, since both the Sun and HP JVMs have too big of a footprint to be used in an embedded environment.
True, but a lot is going on in Java land, and recently it has become fun for WIND. WIND decided from the get go to play it straight and official with Java, and worked with Sun and JavaSoft to deploy Java. Oracle/NCI was a major meeting ground these movers. Meanwhile, INTS jumped ship early on and went with a Java clone, being turned off by Sun's slowness to address numerous problems of embedded Java, and Sun's lack of reality about embedded business in general. Microware established a sweetheart deal with Sun before JavaSoft took over and started creating Java categories.
Sun then acquired Chorus, a move berated on this thread. You might recall William Sheppard from Javasoft posted what amounted to apologies on this thread, and reinforced the importance to Sun of all the RTOS vendors wanting to overlay their RTOS' with embedded Java. Apparently more than an apology on this thread was needed, because all the remaining RTOS vendors jumped ship at HP's first announcement of a reasonably priced, efficient alternative. That is, all but WIND. Sun desperately needs WIND to remain in the fold, causing WIND's relationship with Sun to change overnight. Somehow Sun is going to have to downplay JavaOS and Chorus, and completely redefine the royalty model to keep WIND in the fold. How this all plays out is hard to say, but one thing is clear: WIND is in the driver's seat on Embedded Java and gets to call all the shots.
As for Chorus, remember a year or so ago Chorus/Lucent announced that Chorus was a major RTOS provider to the telecommunications equipment giant? Since then Sun acquired Chorus, and now we see WIND announcing a strategic relationship with Lucent. Obviously Lucent has little interest in becoming dependent on a piece of a division in a hardware company, nor does Sun have much of a commitment in expanding the pure RTOS capability it acquired. If Sun didn't feel this way last year, there is no question but that it does now.
Finally, Ron hemmed and hawed when asked about the decline in deferred revenue, but if I remember correctly, he blamed it on an unnamed Japanese company.
Actually it was Dick Kraber who addressed deferred revenue. He blamed it on accounting transactions with a Japanese distributor.
This is how I understand his explanation. For all those people who haven't noticed, many Asian countries are not exactly open to imports. To overcome trade barriers, in many countries it has been necessary to sell through connected middlemen, or agents. These "channels" buy "inventory" at a discount that they replenish usually on some periodic, established schedule. Occasionally, the replenishment is excessive and requires an adjustment. With WIND, this last happened in 1993 immediately after the IPO, and was a significant factor in WIND's poor FY 1994 performance relative to FY 1993. Since then the potential for allowing this problem to sour the numbers has been contained, no doubt by minimizing inventory sold through these channels. Even so, the nature of the Asian business environment is such that obviously it cannot yet be eliminated completely.
Allen
|
