TO ALL --- AEA 1997 notes
Last week I attended the "American Electronics Association Classic" financial conference and had the opportunity to meet with 25 of the over 400 companies participating. Several companies associated with the embedded computer industry participated but only two, WIND and INTS, were RTOS/tools companies. MWAR did not participate this year, understandably, given the difficulties which they are enduring.
INTEGRATED SYSTEMS, INC. (INTS)
Before discussing Wind River let's first cover the presentation by Integrated Systems. David St. Charles, CEO, made the following comments.
ú Windows CE is not a threat since it is being positioned at the high end, i.e. PCA's, hand held computers, NC's, etc., while INTS is more focused on infrastructure and high volume products.
ú INTS's pRISM+ wins in head-to-head comparisons with WIND's Tornado in independent tests. Also, Tornado does not have an extensible product. Some current Tornado users are switching to pRISM+ despite heavy discounting by Wind River. The average selling price for pRISM+ is $26 to 30K. Hundreds of pRISM+ product have been shipped. Right now INTS is concentrating on building pRISM+ "visibility" so do not look for a significant Q3 or Q4 revenue impact but there will be a big ramp up next fiscal year.
ú I2O has not taken the world by storm. Question whether I2O will find wide acceptance.
ú INTS dominates the digital set top box market.
ú The automotive market will not contribute significantly to revenue before year 2000.
ú Run-time license royalties are stable and run from 10 to 15% of revenue depending upon quarter.
ú INTS has a "hockey stick" with 50% of revenue coming in the last month of the quarter.
ú INTS has about 5000 customers.
ú Q2 service revenue was inflated by $3M due to a one-time contract.
ú FY 98 estimated gross margins are 63% for product and 37% for services.
Here are my impressions for what they are worth. The tone of the presentation was very defensive. The emphasis throughout the presentation was not on performance goals but rather on "building visibility". The frequent references to Wind River left me wondering whether knocking down Wind River was more important than promoting INTS. This more than anything revealed how much they are being hurt by WIND. I asked about the independent comparisons of pRISM+ vs Tornado and where I might get a copy of the studies and who paid for them. The answer is that the tests were commissioned by INTS and have not been published and are not available. I asked if they were forfeiting market share by tying pRISM+ to CORBA. Mr. St. Charles said that he saw CORBA as being the wave of the future not D-COM. Their head of marketing then chimed in with a throw away statement that D-COM can be translated into CORBA therefore they are not precluded from selling into the D_COM market. As for the heavy discounting, WIND's ASP is $50K versus INTS $26-30K. Their stance on I2O was particularly interesting given their press release a month ago promoting their strong offering versus the currently available "rudimentary IRTOS functionality". They seemed to have no interest in talking about their I2O initiative. I left the meeting with the feeling that INTS is a very troubled company. Coupling this with an apparent decision a couple of quarters ago to stop increasing both marketing and R&D spending in order to protect the bottom line leads me to believe that INTS is caught up in a downward spiral.
WIND RIVER SYSTEMS
The presenters for WIND were Ron Abelmann (CEO) and Richard Kraber (CFO). A great amount of information was presented and I went back for a second session to try to capture it all and have additional questions answered. In order to make this more digestible I have broken the material into a series of topics.
FINANCIALS
ú 65% domestic and 35% overseas revenue.
ú Revenue is seasonal: Q1=20% Q2=24% Q3=26% Q4=30%
ú Gross margin = 82%
ú Operating margin = 25% and should rise every year! I asked about the large difference in operating margins between domestic and overseas subsidiaries. It is not meaningful and is just a function of transfer pricing which tries to retain earnings in the US (to the extent allowed by law) in order to minimize overall taxes. Domestic and overseas profitability are no different.
ú Cash and investments (including Long Term) = $241M - will be used to purchase real estate, acquire complementary technology and take small stakes (under 20%) in technology companies.
ú LT debt (conv sub notes) = $140M
ú Company will be buying $2.5M worth of stock per qtr, but this will not completely cover employee options.
ú R. Kraber said that he does not anticipate any further equity or debt offerings.
ú Tornado Avg. Selling Price = $50,000
ú Biggest customers are HP (4%) and Nortel (4%).
ú Revenues have been growing in recent quarters at about 45%
ú Bookings have been growing at over 50%
ú WIND has an internal goal of maintaining a 90-day backlog (guaranteed) so that the predicted numbers are in-hand on the first day of the quarter. This goal has been met for the past few quarters! Only a small portion of this backlog is included in the deferred revenue account on the balance sheet. Most of the backlog consists of orders received but not booked and is thus not visible in the numbers.
ú Ron says that he has visibility 4 to 5 months ahead.
ú WIND has exceeded all analyst's estimates for 14 quarters in a row.
ú Q3 FY98 is solid -on plan- bookings outstanding and a good pipeline.
MARKET SHARE
ú 98% of WIND's business is 32 bit processor based
ú 32 bit is 4% of embedded market
ú 32 bit embedded processor growth is 50 to 100% annually
ú 32 bit RTOS/Tool market is growing 35% annually
ú WIND has approximately 35% market share of the commercial RTOS/tools market and its share is growing.
ú WIND supports all popular 32 bit processors (over 20, far more than anyone else). Note: Prior to 1994 VxWorks was ported free of charge to new processors. Porting takes 9 months and costs $200K to 300k. Ron changed this when he arrived in 1994 so that most porting is now funded yielding gross margins of 60%. If the processor is a success (i.e. sells over 4 million units) then WIND rebates 5% per quarter for up to 10 quarters to the semiconductor company.
GROWTH DRIVERS
ú Availability of 32 bit processors for $10 - 20
ú Complexity of 32 bit apps plus shortage of embedded engineers drives outsourcing
ú WIND has service relationships with semiconductor companies that no one else has. NOTE: A new dimension to these relationships has developed in the last year or so. Prior to the IxWorks distribution agreement with Intel the semiconductor companies would not even discuss the possibility of distributing a RTOS with their chips. They considered it to be important to "keep a level playing field" for all RTOS companies. This attitude seems to be changing.
ú Focus upon telecoms infrastructure - high growth
ú The key to success is not the RTOS. It is the TOOLS and CUSTOMER SUPPORT which provide "time-to-market".
TORNADO
ú Continuing very rapid growth - over 15,000 users currently
ú Over 140 partners
ú UNIX 60% and Windows 95/NT 40%, Windows growing faster than UNIX
ú At least two years ahead of competition
HIRING
ú Low employee turnover - approximately 7% annually
ú Hiring is not a problem despite high standards. They are on schedule for all hiring. WIND's success draws candidates. They also go to where the engineers are - France, Israel, Japan, etc.
ú Customer support is considered to be of the utmost importance. 25% of engineering time is budgeted for this. Customer service personnel and field application engineers comprise 20% of staff. Despite this, customer support is considered to be WIND's weakest area. Ron said that they have to be paranoid about customer support and will be spending heavily on customer support and applications consulting to meet customer needs.
MARKET BREAKDOWN
Wind River has little exposure to the consumer market. Its focus is on infrastructure.
ú Telecom/Datacom/Internet 25-30%
ú Office Automation 20-25%
ú Industrial/Medical 20-25%
ú Military/Space 20-25% (WIND dominates this market)
ú Consumer 5-10%
I2O
Intel will ship a few hundred thousand i960RD/RP chips in the fourth quarter of 1997. This will ramp up during 1998. Intel's estimate is that 10 million of these chips will be shipped in the first 24 months starting this quarter. I asked Ron if all Intel motherboards are now shipping with i960 chips. The answer was yes. Intel has 60 days from the end of the quarter to pay WIND so royalties for this quarter will show up in WIND's Q1 FY99. Three IxWorks distribution deals have been signed - Intel, DEC (who's semiconductor division is now being acquired by Intel) and a third company which has not yet been announced. I asked Ron about Integrated Systems entry into the I2O business. His response was that, although INTS complies with the I2O standard, their implementation is incomplete. He said that there are discussions within WIND on whether or not to offer a scaled down version of IxWorks to compete directly with INTS but that no decision has been made.
DSP
The discussion of Tornado for DSP provided a reality check for me. When WIND's press release on Tornado for DSP began with the phrase "major announcement" I jumped to the conclusion that we were hearing about something of significant financial importance for the near future. It turns out that the importance is of a strategic nature rather than something which will directly contribute significant revenues anytime soon. The strategic importance relates to the expectation that we will increasingly see DSP and microprocessor chips on the same board and Tornado can provide a common programming interface for both devices. Although shipments of DSP chips are expected to outpace CPUs and MPUs for the foreseeable future the need for higher level tools and a commercial RTOS is not as compelling. Also, royalty fees will be significantly lower, perhaps half as much. Ron's estimates that DSP might contribute $1 - 2M to revenue in FY98 and $4 - 5M in FY99. Perhaps in 3 to 5 years DSP revenue could be significant. Perhaps in 10 years it could be 25 % of the buiness. WIND's major competitor in this area is Spectron, a subsidiary of Dialogics. They do about $5M a year in DSP business. WIND is currently in negotiation with TI about porting Tornado to their DSP chips.
EMBEDDED INTERNET DEVICES
I found out that EIDs are a difficult subject to discuss. They cover such a broad spectrum of physical devices (many yet to be conceived) and overlap so much with other categories that it is difficult to become quantitative. The opportunities are large but this area will be more competitive than the infrastructure area that WIND focuses on today.
NETWORK COMPUTERS
Last year at the AEA Classic Ron spoke enthusiastically about Network Computers and predicted that by this time a strong ramp would be starting and that this would produce a big revenue stream for WIND. His frustration was on display this year. He did not expect a significant volume of NCs any time soon, perhaps in another year. I asked him if WIND's operating system was now being used by NCI. The answer was no. In addition, he said that Oracle may design their own.
AUTOMOTIVE
I asked for Ron's views on when the automotive market may become a meaningful source of revenue. He said that there is very strong resistance to the run-time royalty model in the automotive industry. This is not the way that they do business and WIND is not having success changing that. Ron added that without the royalty stream he is not interested in the business. He is pursuing a strategy to hopefully resolve this stalemate. WIND is having discussions with the semiconductor vendors (e.g. Siemens, Motorola, etc.) to the automotive industry. He is trying to establish agreements with them to distribute VxWorks on their chips so that the royalty becomes part of the piecepart price like the I2O arrangement with Intel. He hopes that this will crystallize in the next six months.
QUALCOM
I asked about the nature of WIND's relationship with Qualcom and CDMA. He seemed puzzled why I asked. He said that there is nothing special about the relationship, just another customer, not a big deal. WIND participates in the build out of the ground station and satellite infrastructure. He does not think that this gives them any advantage in the follow-on handset business.
COMPETITION
I asked Ron for his perspective on Integrated System's pRISM+. He believes that INTS was caught completely by surprise two years ago by the Tornado introduction which at that time already had 40 man-years of development. In order to get a competing product on the market they went outside to obtain core technology, a very undesirable situation. He considers the resulting patchwork to be weak and at least two years behind Tornado which WIND continues to invest heavily in.
Someone asked about competition from the Java OS and "Java on the Metal". He thinks that the Java OS has serious limitations for the embedded market. It is much too big and too slow.
I asked Ron his opinion of Windows CE. His view is that MS has some problems to overcome in penetrating the embedded world with Windows CE. First, MS is a channel distributor and the embedded world requires direct selling. Also, CE is not real-time. It has latency times in milliseconds when microseconds are required. MS is also expecting royalties of $15 to $20 per unit which is several times what WIND charges.
Ron said that he does not worry about competition, just Wind River's execution. |
