Sun's Acquisitions Boost Revenues, But Profits Still Elude
Published: January 24, 2006
by Timothy Prickett Morgan
While there are plenty of skeptics concerning Sun Microsystems' acquisition of StorageTek last year for $4.1 billion in cash, that deal, and to a very small extent the acquisition of SeeBeyond, did accomplish one thing: Getting Sun's revenues growing again. Thanks predominantly to those deals, Sun was able to report a sizable revenue increase of 17 percent in the second quarter of fiscal 2006, hitting $3.34 billion in sales compared to $2.84 billion this time last year.
However, shortages of UltraSparc-IV+ systems held down sales, and restructuring and acquisition costs forced Sun to report a loss of $223 million, compared to a tiny $4 million profit it posted in the second quarter of fiscal 2005. In explaining the loss, Sun said that it included $145 million of accounting adjustments relating to the StorageTek and SeeBeyond deals, $10 million in restructuring charges, a $14 million gain on equity investments, and a $3 million tax benefit. Sun also put $55 million in stock-based compensation on the books, too, which reduced net income.
In a conference call with Wall Street analysts, Sun's chairman and chief executive officer, Scott McNealy, said that the company exited the quarter with a backlog of UltraSparc-IV+ server sales of about $100 million--something that no vendor likes admitting unless they are bragging about having the largest sales backlog and pipeline it has seen in a few years, as Sun was. "We just underestimated how popular these UltraSparc-IV+ products would be," explained McNealy, adding that Texas Instruments had executed to Sun's orders, that "yields were fine, and we just didn't order enough." He said that by the end of this quarter or early in the next, the lead time for shipments of UltraSparc-IV+ servers would shrink to its normal one to two weeks for entry servers and two to four weeks for bigger boxes. He would not say how long the lead times were during the fiscal second quarter for Sun Fire servers based on the UltraSparc-IV+ chips. These are Sun's second generation of dual-core processors, and they significantly increase the performance and bang for the buck of the Sparc-based Sun Fire server line, putting them ahead of Hewlett-Packard's HP-UX Integrity machines and a lot closer than past Sparc boxes to IBM's AIX p5 machines.
Ironically, this seems to imply that had Sun not done these acquisitions and had not underestimated demand for its own products, it might actually have had something closer to a breakeven quarter. Sun paid a lot of money for those 1,000 StorageTek sales people, and they had better start peddling a lot of storage and get more other Sun products into accounts if Sun ever hopes to get a decent return on investment for that $4.1 billion in hard, cold cash.
The fact that Wall Street was expecting Sun to have a loss of about 1 cent a share (based on surveys by Thomson/First Call), compared to the 7 cents it reported, and sales of $3.49 billion, compared to the $3.34 billion it reported is problematic for investors of Sun, of course. But, then again, Sun's top brass has been careful in the past several years to set no expectations of its own. Sun plots out a strategic course, explains its rationale and how it hopes to make money, and rolls the dice with the rest of us. It may be an unconventional approach with dealing with a pesky investment community, but it has a certain refreshing honesty to it as well. Wall Street wants vendors to make promises, of course, so it doesn't have to. If investors are annoyed, they ain't got nothing on the analysts at the major brokerage houses that peddle Sun stock and have to call that stock up or down or sideways as we go through each quarter. They don't want to make those calls, and McNealy and Co. has made them do it since business went sour in 2001.
Steve McGowan, Sun's chief financial officer, said in the call that Sun booked $1.44 billion in computer product sales (down 5 percent), $670 million in data product sales (including StorageTek, and obviously up by a lot, but Sun didn't say by how much since it doesn't want anyone to be able to extract StorageTek sales from this unit), and $1.23 billion in services sales (up 23 percent). But Sun did provide enough data to do some back-of-the-envelope estimating. Sun said that overall products sales were up 15 percent year-over-year, and when you do the math backward, that means computer products was about $1.52 billion a year ago and data products was about $310 million. So StorageTek plus whatever organic growth Sun got from its own storage business before the two were merged into the Data Products group brought about $360 million extra dough into this quarter, more than doubling that business. McGowan said that StorageTek contributed $228 million to Sun's services area during the fiscal second quarter, too.
While Sun saw an 87 percent growth rate on its Opteron-based servers, shipping nearly 20,000 units in the quarter, Sun's overall server shipment rate declined by 6 percent to 83,000 units. This means that Sparc-based server sales declined from about 77,600 units a year ago to just over 63,000 units in fiscal Q2--a drop of 19 percent. It is hard to say if the $100 million backlog at the end of the quarter represented all of this approximately 14,600 in Sparc unit shipment declines. If you divide that out, you get an average selling price (ASP) of $6,850 per server. McGowan did say that 75 percent of Sun's server revenue in the quarter came from machines in the one-way to eight-way market, where ASPs are low. But it is hard to believe that even half of that Sparc shipment decline is due to the inability to get UltraSparc-IV+ parts. Sun did not elaborate on this, but merely said that it saw revenue growth in Opteron-based servers and midrange and high-end storage arrays.
While Sun does not yet break out its software revenues as a separate category, it did say that it added 148,000 subscriptions to its Java Enterprise System middleware, pushing the total seat count up to over 1.1 million seats thanks in large part to a big deal it signed with American Express. McNealy also said that Sun exited the second fiscal quarter with about 3.7 million registered downloads for the Solaris 10 operating system and said that Sun reckoned as of today, which is close to the one-year anniversary of the launch of Solaris 10, it was close to 4 million cumulative units shipped. While Sun has not talked about how much money Solaris 10 has generated directly, the company says that Solaris training sales are rising, which it views as a leading indicator. "Our strategy of free is working," explained McNealy.
A few of the analysts on the call pressed Sun to explain how the new "Niagara" Sparc T1 servers and the Opteron-based "Galaxy" servers were cannibalizing the current Sun installed base, and both McNealy and Jonathan Schwartz, Sun's president and chief operating officer, simultaneously chimed in that "we have already been pre-cannibalized," meaning that alternative Unixes--predominantly IBM's AIX/Power combo but also Lintel platforms--had spent years attacking the low-end of the Sparc line where the Sun Fire machines using the T1s and the Opterons are now aimed. They said that these machines were not designed to replace existing Sparc machines so much as allow Sun to compete against Windows and Linux on X64 processors from Intel and AMD and to keep the remaining Unix competitors at bay.
Schwartz said that about one-third of the Galaxy boxes going out the door were configured with Solaris--up from practically none at the beginning of the Galaxy launch--with the remainder running Linux, Windows, or having no operating system when they shipped. "We're seeing good growth of Solaris on Opteron," he said. McNealy went a little further and said that the barrier to exit from Solaris to Linux was pretty low over the past several years, because of the commonality of Linux and Unix. But that with the Linux runtime environment that Sun has created to run on its Opteron-based servers, he said that "the barrier to exit from Linux back to Solaris is even lower." Sun is happy to sell Galaxy machines with Linux on them, and is working to get Linux and BSD running on its T1 processors as well with the communities that steer their development. But clearly, Sun wants customers using its servers to prefer Solaris and then pay it for support and services.