Sun in Transition As Announcements Loom

by Timothy Prickett Morgan

Next Tuesday, Sun Microsystems will roll out the next batch of new products and updates as part of its "Network Computing 04Q2" quarterly product announcements. While Sun's customers, partners, and competitors are always interested in seeing what Sun is doing and where it is going, this announcement will be a little bit different, if the word on the street is right. You see, this will be the first time that Jonathan Schwartz, Sun's new president and chief operating officer, will be running the show.

Schwartz was given his titles and power only a few weeks ago. After running Sun's software business for a while, Schwartz comes on board as the company is trying to make the transition to UltraSparc-IV, is readying a new line of Opteron-based entry and midrange servers, is pushing to get Solaris 10 to market by the fall, and is facing threats from Unix competitors, Windows servers, and Linux alternatives. With Sun trying to get back to profitability, all eyes will be on Schwartz. This has to be a bit comforting for Scott McNealy, one of Sun's founders and the CEO and chairman who has been on the hot seat for the past two years as Sun has struggled with the realities of a new IT market and changed just about all of its top brass. With Sun having recently settled its litigation with bitter rival Microsoft, McNealy is probably not going to be cracking a lot of jokes about Windows, but he may be in a more chipper mood, particularly since this is the tail end of Sun's fiscal 2004 fourth quarter.

Sun has ceased giving mid-quarter updates to Wall Street analysts, but Steve Milunovich, chief IT analyst at Merrill Lynch, this week released a research report based on his own conversations with Sun's customers and partners. Milunovich said that he believed demand for Sun's products is improving but that supply shortages could hamper sales. June is going to be an intense month for Sun, particularly since about 40 percent of the sales Sun makes in the fourth quarter happen in those four weeks that span Memorial Day to Independence Day in the American mindset. Milunovich is estimating that Sun will have overall sales of $2.9 billion in the fiscal fourth quarter, down 2 percent from last year but up around 11 percent from the fiscal third quarter. He also expects that Sun will lose around 3 cents a share. Thomson/First Call puts the street consensus at $2.88 billion in sales and a loss of 4 cents a share.

Milunovich said that Sun's biggest customers in the financial and telecommunications sectors are starting to spend again, echoing comments made in recent weeks by Sun executives. Milunovich singled out the high-end Sun Fire 15000 as a machine with good interest, but said that Sun's home-grown midrange StorEdge storage arrays might not be selling well and that Sun's overall storage sales might be hampered by the company's lack of a network-attached storage (NAS) array product. If Solaris is so good at managing storage, and can run on X86 iron, it is a mystery why Sun hasn't created its own NAS box running on cheap Serial ATA or SCSI drives. Perhaps this is in the cards.

Even before Schwartz took over, all of Sun's executives were looking around, trying to figure out how to switch from the product push-sell cycle to utility-style pricing with recurring revenues. There are a lot of reasons why Schwartz ascended to the president and COO position, and one of them is that he understands what it means to do this, because he has been selling software this way. The Java Enterprise System, which costs $100 per employee per month and was announced only a few months ago, is a harbinger of the kind of simple, direct sales approach Sun will undoubtedly try to bring to the server, storage, and services markets. Milunovich also said that Sun's Java Enterprise System, which includes a full suite of middleware (except a database), is priced aggressively but that large customers tend to want to buy best-of-breed and mix and match components. Many customers just don't get it yet, and Sun has a lot of work to make this new pricing model palatable and to get enough customers so that it feels like a safe option.

If Sun wants to be an IT player from here on out, it has to make the new pricing models work--which are disruptive to revenue and profit streams--and do all the technical feats that it has promised, such as getting truly innovative and scalable Opteron machines to market, demonstrating the benefits of its future "Niagara" and "Rock" multicore, multithreaded processors, and continuing to sell its UltraSparc-IV and UltraSparc-IV+ servers against very competitive Unix offerings from IBM, Hewlett-Packard, and Fujitsu-Siemens.

Sun has to do all of this while cutting costs even further than it already has. Milunovich's surveys of chief information officers indicate that two-thirds of those polled doubt that Sun can recover. They have good reason to doubt Sun, since execution is the real issue. But only a fool would count Sun out in 2004, just like only a fool would have counted out IBM in 1994, when it was on the ropes and looking pretty dead. Make no mistake about it: Sun is energized, and it is thinking outside of the box, placing bets, and shaking things up. This is a good start toward its own recovery as the IT spending environment itself recovers.

More than anything else, Sun has to stop being vague about future server platforms and start making commitments and roadmaps, just as it has been doing with Solaris 10 for the past year. Schwartz would be wise to start talking, since the lack of clarity on future Sparc and Opteron servers is making customers jumpy. Next week is a good time to start talking, before the customers balk and walk.