Does this imply that you expect WIND to begin beating earnings estimates by more than the 1 or 2 cents per quarter that we've grown accustomed to? When would you expect that to begin?
There is enormous confusion about WIND beating the estimates by 2 cents. That rule only applies to the upcoming quarter. Analysts estimates for the year, and certainly for the following year always understate performance more than by 2 cents per quarter, or 8 cents per year. Today analysts seem to be projecting about $1.12 for FY 2000-two years from now. I think that number is a silly understatement of WIND's expected performance over the next two years. My estimate is around $1.64 in FY 2000, which equates to a share price of $107 with a multiple of 65. The rich 65 multiple should be minimal at that time, given WIND's consistent string of 24 quarters beating analysts' estimates, consistent EPS growth rate in excess of 50, and extremely favorable outlook. So, the triple results from better EPS performance than estimated by analysts, and a slight multiple improvement-more in line with WIND's average price performance.
Believe it or not, my estimates assume a modest contribution to revenues from I2O, and revenue growth of only 45% in FY 1999 and 46% in FY 2000. Meanwhile, EPS for the two years is projected to grow at 61% and 56%, respectively. (I even assume that Dick Kraber gets additional people!)
But anybody can make wild claims about WIND performing far in excess of analysts' estimates. What makes my estimates worth considering? What about I2O?
Let's start by looking closer at I2O. Neither Intel, WIND nor I can project I2O unit volume or revenues downstream with any confidence whatsoever. Anybody that looks carefully at the problem sees the same thing: annual volume of 100 million PCs, 2.2 million NT Workstations and 2 million NT servers this year. Unit volume of PCs is growing at about 16% annually; workstations 45% and servers 38%. Each future server can be equipped with anywhere from one to 16 I2O chips. Server devices, such as disk arrays, NICs and tape backup systems, may themselves be equipped with I2O processors. Nobody knows when I2O will migrate across the entire line of Intel architecture computers and devices. Consequently, I assume that within two years only a small part of the potential I2O market is penetrated: 1% for PCs, 10% for workstations and 25% for servers, yielding I2O royalties in two years of $6.3 million, versus the market potential of about $200 million.
As great as I2O is, and I am constantly flabbergasted at its potential, there are about 80 I2O design wins. Remember when the thread discussed extravagant statements by QNX management bragging about their potential exponential revenue growth based on 200 design wins. (Indeed, if I2O has the incredible potential we perceive based on 80 design wins, 200 design wins must equate easily to the $200 million revenue predicted by QNX management. Note that QNX is private, so their management can say whatever they think will happen, rather than what they absolutely know they can accomplish.) QNX sees their design wins propelling them past most RTOS vendors, and up to the level of WIND.
I2O is great, but what about WIND's other design wins? How do they compare to QNX? How dependent is WIND on I2O?
Now, think about what I am about to say very carefully. Measured in terms of design wins, I2O comprises only about 1% of WIND's business. Today, WIND probably has in excess of 6,000 design wins. Anything I have ever said about I2O is an understatement; yet I2O is hardly a blip on WIND's radar screen.
Just like QNX, not all of those design wins will get to market. Many will get to market, but with relatively small production runs. Many will get to market with sizable production runs. And a significant number will get to market with huge production runs. My non-I2O estimates are based on statistical extrapolations of revenues and costs, which under the circumstances are credible, irrespective of what happens in Asia, in the military, in automobiles, in desktop boxes, in digital cameras, whether JetSend takes off, whether cable modems prove a success, or even whether El Nino continues raining on California. Six thousand plus design wins, across every conceivable walk of life, has less risk mathematically than U.S. Treasuries.
What are some of the other markets WIND has penetrated that accounts for some of the 6,000 design wins?
I doubt there is a telephony or network equipment manufacturer with which WIND is not doing business. Almost a year ago in a conference, Ron Abelmann said you can't make a telephone call without involving VxWorks. During the last conference call, Jerry Fiddler indicated that VxWorks is in virtually every cable modem, and will be in about half of all xDSL modems that will be developed. With HP, ADBE and EFII, we that we know VxWorks will be showing up in most printers. WIND dominates military systems, which is scheduled to skyrocket in production volumes. Everybody (WIND management hinted, and analysts have parroted) expects significant announcements about deals in the automobile sector. JetSend, digital cameras, auto navigation, lottery terminals, etc. etc. Add it all up and there are more than 6,000 individual products that may be constructed on top of VxWorks.
Despite the awesome potential of all those design wins, my projections are based on straightforward statistical extrapolation. I suspect that WIND management is sufficiently conservative that my numbers will be easily met, but not exceeded by much, if at all. Instead, two years from now, we will be sitting on $1.64 in per share earnings discussing why 20,000 design wins easily justifies 50% EPS growth over the next two years.
Allen |
