>although it [WINDs qtr] is over a month away - do you have any feel for the numbers they will report and the date which they will report?...(particularly given the latest reports from MWAR and INTS)
First, let me assure you that I believe WIND will meet its first quarter expectations of $.09 per share versus $.06 per share last year. In fact, I have difficulty believing WIND will produce EPS less than $.13.
That being said, let's admit that my answer is not very satisfying. Deep down, as a microcosm of the market itself, you fear that WIND will follow INTS and MWAR with disappointing performance. How is it possible that INTS is undeniably struggling and MWAR is about to yell uncle, while WIND can be breezing along increasing both revenues and earnings like its a walk in the park? Sure, you have read most of the posts on this thread over the last few months, but the fear remains.
The financial market encounters unexpected sector under-performances all the time, and is constantly alert for this possibility. It happened in 1995/1996 in semiconductors, after pundits promised that the semiconductor business cycles would be flattened by ever-increasing demand. It happened in 1996 in the digital broadcasting/media tools sector when companies like Avid Technology, Discreet Logic, Accom, Scitex, Data Translations, Pinnacle Systems and Sonic Solutions all hit unexpected stone walls, with signs of life just beginning to be noticeable well into 1997. When this occurs all stocks in the sector tend to perform poorly for a period of time, and get duly punished by the market. You may be excused if you would like not to be long on any sector during such tumultuous times.
Is some kind of a sector slowdown happening today in the embedded systems space? Is this the real explanation for INTS and MWARs troubles? To answer this question, you begin by trying to understand why sector slowdowns happen. What happened in semiconductors? What went wrong in broadcasting/digital media? Are there any parallels in embedded systems?
It should take you about one minute to conclude that the semiconductor business cycles bears no direct relation to the embedded systems space, and the manufacturing over-capacity that developed in that sector has no current parallel for embedded systems.
However, what went wrong with the broadcasting/digital media tools sector makes for interesting analysis and parallels. In my mind, it all boils down to the timing of an industry-wide shift of enormous magnitude from analog to digital. The forces of this shift are so huge that they can shake the broadcasting/media industry to its very core. During the first half of this decade, early adopters gobbled up all the emerging digital capability they could lay their experimental hands on, showing insatiable appetites for creative uses of digital technology. Unfortunately for all the companies doing business in the sector, rapid growth came to a screeching halt in late 1995 when the costs, difficulties and confusion associated with a complete digital conversion were not breached by any compelling reasons to take the plunge. HDTV, DVD, Satellite TV, VOD and the beginnings of serious commitment by major broadcasting entities are beginning only now to provide the reasons, and if so, you can expect this sector to show accelerating growth throughout the remainder of the decade.
What is particularly interesting about what happened to broadcasting/digital media is the force and magnitude of the hiatus that developed. This causes me to conclude that the greater and more difficult is the infrastructure decisions that need to be made to fully deploy the emerging technology, the greater will be the hiatus that can be expected. Make the decision too momentous, and deployment might never happen.
You could argue that embedded systems tools vendors face a similar marketing dynamic to that faced by the broadcasting/media tools vendors. Emerging technologies often face one of two common forms of infrastructure obstacles that must be overcome in order to reap the rewards of mainstream sales. Either an existing infrastructure must be replaced, like replacing analog with digital in broadcasting, or new infrastructure must be built, like upgrading services provided by Cable TV companies to include VOD. In the former situation, the decision-maker is slowed down by the enormity and expense of throwing out existing infrastructure and converting to a new business paradigm. In the latter situation, the decision-maker must invest hugely in new services of unknown value to the customer. Both, very risky decisions requiring absolutely compelling reasons before the decision-makers will go forward.
The market may be seeing these kinds of possibilities and parallels, and they may well explain MWARs troubles, for example. The STB market niche started as a boon to product license sales for MWAR, but then went into suspended animation awaiting the big infrastructure decisions concerning deployment. While the broadcasting/media toolmakers got hung up by the enormity of infrastructure replacement decisions, MWAR is being held hostage to the enormity and risk of infrastructure investments required to provide VOD. MWAR reloaded and pointed its RTOS at wireless and internet appliance niches, which they can't hope to dominate like they did early in the STB niche.
INTS is in many more market niches than MWAR, and there is evidence that their RTOS sales are growing reasonably, hardly a sector slowdown. As a late-comer to RTOS, INTS has made bold but risky moves to obtain a competitive toolset, which has led to on-going problems. Integrating all their acquisitions has proven costly, both in terms of expenses and lost revenues. (Epilogue's revenues disappointed during the last quarter, they claim, due to its physical move from Albuquerque to California.) pRISM+ may be too little too late, and pSOS has been loosing ground to VxWorks in important markets like military and telecommunications.
WINDs business base spans many market niches, each with its own set of dynamics spanning early (early adopter, experimental, development) versus late (infrastructure) sales. For example, the automobile market breaks down into two sub-markets: (1) under-the-hood and (2) in-the-cab products. Under-the-hood products consist of engine, transmission, brake, battery, etc. controllers. While all the manufacturers use embedded controllers for engine injectors, for example, virtually all of them are home-grown. The infrastructure replacement decision to switch to COTS (commercial off the shelf) is developing very, very slowly in this traditionally conservative niche. If we need this business to make WINDs numbers this quarter, I would be as concerned as anyone.
Automobile in-the-cab products include automobile navigation, on-board computers, hands-off phones, wireless emergency and other roadside services, and many more products on the drawing board, such as anti-collision systems. Of all these, automobile navigation may be most advanced in terms of life cycle dynamics, slowed mostly by lack of detailed maps and marketing, although WIND should be noticing significant revenues from the Hitachi-Nissan joint venture to push the Birdview navigation system. The phones and on-board computers have yet to advance to a stage requiring CTOS. Wireless emergency and roadside services will explode when low-elevation satellites become available, and once infrastructure investments to track available services and set up message centers have been completed. With all its opportunity, nevertheless the automobile niche is remains in the experimental stage, contributing only slightly to current revenues for commercial RTOS vendors. When INTS advertises its strategic relationship with Mercedes-Benz, for example, they often fail to emphasize the relationship is in regard to MatrixX, not pSOS.
The ultimate experiment with an RTOS is the on-going Pathfinder to Mars. VxWorks is controlling the cruise adjustments. Once there, the rover can thank VxWorks for its ability to explore even a limited portion of the Red Planet. Message: COTS works for mission critical applications - at a fraction of the time and cost required by predecessor MILspec or NASAspec products. I think the military is going full-speed-ahead replacing existing infrastructure with CTOS-based enhancements, as well as developing scads of new and exciting applications. Funded by order-of-magnitude savings derived from CTOS, the military budget is practically unlimited for building these applications - despite the appearance of a shrinking military budget. The new technology also reduces manpower requirements, thereby providing even a larger share of the pie for embedded applications. WIND is king-of-the-mountain in VMEbus military applications, which is starting to payoff surprisingly well, as Ron Abelmann indicated in a few months ago in an analyst conference call. These revenues would stem primarily from product license fees as untold numbers of contractors pay to get operational, but increasingly also from run-time licenses as many products now are promoted to full-scale operation, and many, many others stall at small but high-cost-per-unit run-time numbers. Motorola's commission to WIND to develop a transition kit that enables an application developed in pSOS or VMEexec (Motorola's VME RTOS) to be moved to VxWorks is suggestive of the extent of WIND's dominance in the military arena.
Office Automation is late stage in HP. I suspect the company wouldn't even consider developing another electronic gadget (printer, X-terminal, cable modem, testing equipment) with anything but a commercial RTOS. It even appears that they limit their choice nowadays to VxWorks. HP is a leader in consumer electronics, and where it goes others will follow. This makes me think Office Automation should be on the threshold of major expansion for commercial RTOS vendors, with WIND gaining market share because of its toolset and I2O. Most OA manufacturers should be contemplating an I2O front-end to easily make their product network compatible and advanced simultaneously. Whether or not they include I2O immediately, they should have been introduced to Tornado through Intel's I2O promotion, which included a sample CD containing Tornado for IxWorks. Thus, OA product licenses and run-times should be advancing smartly quarter to quarter on a sequential basis.
Telecommunications and networking is so huge a market, with so many new products coming on-line, thanks to the internet/intranet/extranet and normal rapid growth, that it can be booming for the likes of commercial RTOS vendors while still being early stage. If you notice the number of Telcom and Network companies committing to WIND over the last couple of years, it would appear that the decision to switch or upgrade to COTS at least for many new products has been made by many of them. This alone should amount to an expansive business, not only of product licenses but run-times. Meanwhile, many of these companies harbor in-house OS groups that protect mainline products from CTOS invasion. As these groups loose influence with managers concerned with functionality, time-to-market and costs, there should be a rush of new business to commercial RTOS vendors. Finally, Telcom operators are needful of massive infrastructure replacement, from obsolete switched circuits to packet networks, which will occur rapidly and be huge indeed when it starts in earnest. To the extent that the VMEbus is reapplied from military use to telcomm applications, and to the extent that PCI is becoming a technology favorite with its bus-expanding I2O accomplice, WIND is well established to continue benefiting more than other RTOS vendors from this windfall.
Wireless is both early and late stage, depending. The most attractive thing about digital wireless is that it doesn't have any history (read in-house OS which has to be displaced). The infrastructure side of wireless is already firmly in the hands of the RTOS vendors, with WIND well represented if not totally dominate. Of course the big numbers are in the handsets, which I suspect is early stage. I gather manufacturers are just beginning to pack in enough functionality to require a 32-bit processor, but since this is all new, when it comes it will be mostly COTS. With GSM and CDMA PCS systems deploying worldwide as we speak, and Wireless Local Loop and LEO satellites on the way, this niche should develop quickly indeed. In particular, WLL and LEO applications should demand 32-bit processors from the outset. WIND's work with QCOM on the Globalstar infrastructure and handset developments is early but will pay off big in a couple of years, since all the major infrastructure decisions are history. Already, VxWorks is inside the Orbital Sciences satellites, as yet another example.
Internet appliances, like wireless, are completely new and do not require displacing ingrained architectures. Also, internet appliances, particularly wired devices, are easy to build, laid on top of existing internet infrastructure. Most internet appliances do not require Java, or anything fancy like a NC/OS. Consequently, this market should expand explosively from the outset, with virtually no pause from development to full-scale deployment. Needless to say this market benefits tremendously from the wireless infrastructure build-out.
The Network Computer (NC) is early stage and will not return significant revenues until 1998 and beyond. Of course, those returns should be well worth the wait, even through there are lots of infrastructure decisions that must be made by implementing organizations.
I2O is early stage, but all the signs are encouraging that the beginnings of late stage deployment should be noticeable during the second half of this year. Lots of I2O products are beginning to appear. When I2O becomes the standard for enabling inter-computer communication, and easiest and most robust way to add a device to a network, it will become an infrastructure requirement and ramp rapidly. Although my conservative estimate of WIND revenues for the up-coming quarter exclude I2O sales, I believe that I2O alone will contribute $.04 cents to the bottom line this year, and $.20 next year.
Consumer electronics, games, and anything not elsewhere specified (NSE) no doubt contribute to all categories of WIND revenues, but these products tend to be extremely price sensitive and value engineered. I would count on WIND excelling in these categories only because its broad market leadership and toolset maturity gives it a natural preference over adequate alternative choices. In other words, WIND should do fine in this broad category, while conceding many, unseen small niches to the competition. In the final analysis, this category may become the largest in unit and revenue terms.
If you put all these niches together and multiply by 10, 100 or more, you begin to define the embedded systems space. Each niche has its own set of life-cycle dynamics, often completely independent of other niches. The infrastructure barriers and requirements of under-the-hood automobiles is virtually unrelated to military communications and signal processing, or to office automation. More important, unlike the screeching halt that broadcasting/digital media toolmakers experienced, the embedded systems sector is not even nearly homogeneous enough to support expanded business cycles. Each niche, and in some cases sub-niche, gets caught up in its very own set of infrastructure decisions, and rises and falls according to its own distinctive rhythm. Since the long-term trend of each niche favors commercial RTOS vendors, overall the sector should see consistent, unrelenting growth, ultimately stretching across every corner of human endeavor.
WIND spans most of these market niches, and therefore should be a dependable earner for the foreseeable future.
Allen |
