Sun Continues to Transition Products and Lose Money

by Timothy Prickett Morgan

Sun Microsystems announced its financial results for the first quarter of fiscal 2006 this week, and it is safe to presume that Wall Street will be neither surprised nor happy with the results, and that Sun's top brass don't really care all that much what Wall Street thinks because they believe that the complex transformation they are undertaking with Sun is the right course of action for the company.

This is a kind of détente, and it is what happens when a collection of large, slowly moving objects hit a flexible barrier. That's a far cry better than what happens when an infinite force hits an immovable object, but it is a lot less exciting--that's for sure.

Anyone who has invested in Sun wants much quicker results than Sun--or indeed, any other company in the same situation and IT markets--could deliver. The same type of people (and indeed, some of the same people) were just as impatient with the transformation at IBM a decade ago, they gloated over the implosion and explosion of the AT&T-NCR merger, and they for the most part had a strong distaste for the Compaq-Digital and then Hewlett-Packard-Compaq mergers, too. Wall Street is about making money today or in the next quarter or so, and companies are about making money today but staying in business over the long haul. These views can never be rectified, except when a high-flying company like Sun is making money hand-over-fist and growing at 20 percent a year, as it did in the late 1990s. Back them, Wall Street was fawning all over Sun like a fair-weather friend, and analysts and brokers were making tons of money selling Sun stock. Good times, in some ways. But if times hadn't been so good, maybe Sun wouldn't have had the wrong products on the market in 2002, 2003, and 2004. The fact is, the people who work at Sun and the customers who buy its wares want Sun (and all of their IT vendors) to stay in business a long, long time while remaining competitively priced, and they take a longer view. The analysts need only take a walk over to their own IT departments to get a warm cup of reality. Put some sugar in it, because it is a bitter brew. What Sun is trying to do is complex: service old Sparc businesses, start new X64 and middleware businesses, and revitalize its Unix software business while not going broke. Truth be told, it is amazing that the losses are not much higher. If there is any surprise, that's it. And, by the way, a decade ago, I thought the same thing about IBM as Chairman and CEO Lou Gerstner was teaching that blue elephant to dance.

The fiscal first quarter at Sun wasn't great, but it wasn't a disaster. Including the acquisitions of StorageTek and SeeBeyond, which closed in August, Sun was able to generate revenues of $2.73 billion, up 3.7 percent compared to the same quarter a year ago, and because of costs associated with those two acquisitions (which together contributed $65 million in operating losses) and a $50 million cost associated with stock-based compensation, Sun was pushed to a net loss of $123 million, or 4 cents a share, a little bit less than the $133 million loss it booked in the year ago quarter. And while Sun generated $224 million in positive cash flow in the quarter--cash flow being what you talk about when you don't have black ink below the bottom line--later in the call with Wall Street analysts, Steve McGowan made it clear that this $224 million included a $130 million benefit from a tax refund. These numbers include $226 million in revenue, $99 million in gross margin, $17 million in R&D costs, and $87 million in sales, general, and administrative costs from the StorageTek and SeeBeyond acquisitions. While you can complain that Sun might have done something better with $4.8 billion than buy StorageTek, it will take a long time before anyone can say for sure and the deal has not done any damage to Sun's financials--in fact, it help propped them up a bit. Perhaps the most surprising thing is that because of falling component prices and other cost-cutting initiatives, Sun was able to increase gross margin on its revenues in the quarter by 3.3 percent, hitting 44.1 percent.

This is all the more remarkable when you drill down into Sun's product sales for the quarter. The impending launch of the "Galaxy" entry Opteron servers and the "Panther" dual-core UltraSparc-IV+ processors didn't exactly help sales in the quarter. Sun's chairman and CEO, Scott McNealy, and president and chief operating officer, Jonathan Schwartz, said that the Galaxy machines and the Panther chip announcement may have caused some customers to defer purchases. But, then when they were asked about how much of the company's $1.4 billion product backlog was attributable to the Galaxy machines, McGowan said it was "non-material." That's because these machines only cost a few grand and you have to sell a lot of them to show any kind of money at all. Welcome to the server business in 2005, where Moore's Law cuts both ways.

Sun is clearly counting on selling a lot of these Galaxy boxes, which are a lot better than the Sun Fire V20z and V40z machines that the company has been peddling to date using Opteron processors and running Windows, Linux, and Solaris operating systems. Still, even without the Galaxies, Sun was able to sell 14,000 X64 servers in the quarter, an increase of 109 percent in shipments compared to last year, which is pretty impressive when you consider that everyone knew the Galaxies were on the way. Sun shipped a total of 71,000 servers in the quarter, including Opteron and Sparc machines, and total shipments were up 2 percent. What none of the Sun executives said, but which you can figure out when you do some math, is that Sparc-based server shipments fell by 9.3 percent in the quarter, down to 57,000 from 62,900 a year ago. McGowan said a number of times that 77 percent of Sun's server revenues were for machines with one to eight processors, which is an astounding change in mix from a few years ago when Sun was selling baby rack servers and big iron servers like crazy. Revenue growth in the four-way to eight-way server segment was up, as were revenues for X64 servers. But Sun's entire Sparc base is down-shifting to lower-cost, more powerful, dual-core machines, and this is a phenomenon that that all server vendors are going to have to cope with. The amazing thing is that Sun is able to grow gross margins at all in such an environment. But, what counts on Wall Street is real profits, and in the aggregate, it is hard to say if Sun is making any money at all in the server business.

Sun's product revenues were up 1.6 percent to $1.7 billion, but computer product revenues (which includes workstations, servers, storage, and software) fell by 6 percent. Thanks to some help from StorageTek, network storage sales within the computer products group were up 34 percent. Sun cited its midrange and high-end disk arrays as doing well. Sun did not break out software sales as a separate item, but boasted that it did boost the number of Java Enterprise System subscriber seats by 338,000 in the quarter (thanks in large part to a big deal with General Electric), hitting 957,000 cumulative seats licensed. Sun also said that 53 million cumulative StarOffice/OpenOffice.org packages have been downloaded, and reiterated that Solaris 10 registered licenses have crested above 3 million units since shipments began at the end of January. On the services front, sales were up 7 percent to $1.02 billion, with support revenues up 12 percent to $835 million and education and training sales down 10 percent against a tough compared (21 percent growth a year ago) to $187 million.

On a geographical basis, McGowan said that sales in the United States were up 5 percent in the quarter to $1.16 billion, but he said that sales in the financial services, telecommunications, and government sectors where Sun gets a lot of its money, were soft. Sales in Europe were down 1 percent to $912 million, and sales in Japan were down 9 percent to $166 million. The remaining parts of the world accounted for $489 million in sales, up 16 percent.

Sun exited the first fiscal quarter with $4.5 billion in cash and equivalents, with $1.5 billion in net cash.

The question that Wall Street most wanted to have answered was when Sun's core business--that means servers and operating systems--would rebound. McNealy, McGowan, and Schwartz did not--and could not--answer that directly, but they said again and again that the "Niagara" Sparc-based, 32-threaded processor was going to come to market soon and ahead of schedule (probably in two weeks is my guess), that the Panther UltraSparc-IVs were being well received by the Sparc base, and that the Galaxy Opteron servers represented the first time in many years when Sun was absolutely competitive with X86 (and now X64) alternatives. Basically, we will have to wait and see. "We went from nowhere to number six in the X64 server business," said McNealy. "We have our sights set on the number four spot by the end of this year." He must have meant Sun's fiscal 2006 year, because there is no way Sun can do that in the next nine weeks.

In a separate announcement, Sun said that McGowan, who has been at Sun for 14 years and who has been CFO since 2002, will retire at the end of Sun's fiscal 2006 year. The announcement was made last week at Sun's annual shareholder meeting. McGowan, who is only 57, hasn't said what his plans are, but he clearly wants to leave Sun in better shape than when he was handed the financial reins three years ago. There is a decent shot that it could happen, too, if Sun manages its costs and product transitions.