Wall Street Makes IBM, Sun Strange Bedfellows?
Published: March 23, 2009
by Timothy Prickett Morgan
The details are sketchy, and the two companies are not confirming or denying that something is going down, but the Wall Street Journal broke a story last Tuesday, right smack dab in the middle of the opening session of Sun Microsystems' CommunityOne East developer conference in New York, that IBM was in talks to shell out between $7.4 billion and $8.2 billion to acquire Sun. People familiar with the negotiations, which were apparently still going on as I write this on Thursday afternoon, leaked the fact that there were talks to the Associated Press and Reuters.
Personally, I blame Wall Street for this happening now, if a Sun acquisition by IBM does come to pass. (The rumors I hear is the deal will happen early this week, but talks could break down as I write this over the weekend.) But a lot of the blame goes out to the former top brass at Sun--the old guard under Sun founder Scott McNealy--who missed the shift from open systems to open source and let Solaris be surpassed by Linux and who let the also-ran Unix player that IBM was during the dot-com boom get the upper hand and, to put it bluntly, kick the guts out of Sun's formerly triumphant Unix server business during the dot-com bust. Sun has amended its ways, open sourcing its software and embracing x64 servers, but Wall Street has not given Sun time to heal. As far as I am concerned, it is Wall Street that has created the economic mess that has compounded Sun's problems, as well as all IT players, forcing the company's top brass to shop the company around, looking for an arranged marriage--if you believe the press reports. Both IBM and Sun have declined to comment to me about any of this.
The obvious choice to take over Sun is Fujitsu, Sun's existing Sparc partner and the maker of the high-end Sparc64 VII quad-core chips and their related Sparc Enterprise servers that both companies currently sell. Hewlett-Packard is less of a good fit, and that company is digesting Electronic Data Systems and is no doubt conserving cash in this downturn as it cuts pay and makes its own rounds of layoffs. Both HP and Fujitsu were apparently approached by Sun for a merger, and both declined. It is hard to say how many other parties have talked to Sun, but there have probably been others. But this list of companies who can and might want to pay for a $12 billion, non-profitable, IT powerhouse is pretty short.
Just like, I suspect, the attention span of Sun's financial backers must be getting short. Takeover capitalists Kohlberg Kravis Roberts kicked in $700 million in financing back in January 2007 to Sun, when times were still good and we were not in recession. Southeastern Asset Management, an obscure equity firm based in Memphis, has ratcheted up its Sun stake to 22.3 percent through the fall last year and early part of this year. KKR has one seat on the Sun board of directors, and SAM has the right to put two board members on. (In early February, SAM appointed a former CIO of Fannie Mae and Merrill Lynch to the Sun board. It is hard to say how many institutional investors have sizable bits of Sun, but it is not hard to see that if a few major shareholders have run out of patience, they could be pushing Sun's top brass to make a deal they almost certainly do not want to do.
With IBM bragging about how much cash it can generate from its mainframe monopoly and immense services business (which all feed back into those top 5,000 accounts that feed the Big Blue beast), the company is on the shortlist of possible buyers, even if you would have to search far and wide to find two companies least likely to merge. I once joked that the only way I would see Sun and Microsoft bury the hatchet was in each other's heads, and aside from a bit of détente over Java, IBM and Sun have agreed on about as many things as a man and his mother-in-law: They both agree that her daughter is wonderful, but that is about all they concur upon.
There are so many things that are creepy and wrong to me about this deal that it is hard to know where to begin. But I will start here. When he was a young man and on until I was a young boy, my father had a string of American Motors cars, and they were the best cars he ever had. He had some used models when he was young, before my sister and I came along. The year I was born, he spent $3,200 to get a new '65 Rambler Classic sedan, his first new car, which later that year got in an argument with a spare truck tire that fell off a garbage truck on a highway (while my Mom was driving) and lost the argument. (Mom was fine, but still flinches when we talk about it.) He then got a '66 Rambler American, a black one that lives on the edge of my memory, and we got a '68 Rambler American wagon at later. (My Grandpa had a '66 Ambassador, the big eight cylinder model, and Dad still wants that car just because of the way it floated down the road.) When the energy crunch hit, my Dad and Grandpa shifted to Chrysler cars, despite their love of AMC cars, because Chrysler delivered cheap compacts to the market first. The deal was pretty much sealed when my Grandpa became the master mechanic at a huge Chrysler dealership in western Pennsylvania. It has been four decades since that '65 Rambler was smashed up, and my Dad still says no one could beat an American Motors vehicle of that time in terms of bang for the buck, smoother ride, ease of maintenance, and basic engineering.
As it turns out, the vice chairman of American Motors was none other than William McNealy, who also did the company's marketing. Just a few years after his son, Scott, had finished up his MBA at Stanford, American Motors, an amalgam of a number of car makers cobbled together to fight against the Big Three (General Motors, Ford, and Chrysler) in the 1950s and 1960s, was selling itself off to Chrysler when it ran out of cash to innovate. AMC survives as the Jeep brand today. Being number four is not a good place to be.
Sun fought like crazy to engineer great and open workstations in the 1980s and servers in the 1990s, and the fledgling startup got lucky to buy the "Starfire" server line from an absolutely idiotic Silicon Graphics in 1996, giving Sun the big Unix iron to be the "dot in dot-com" when no one else had big Unix iron, making it the number three server maker during that boom time. People talk about those UltraSparc-II machines and Solaris like my Dad talks about that Rambler, and also the way that other people talk about AS/400s and RPG. But Sun has been languishing as the number four server vendor since the dot-com bust. The UltraSparc-III chips and servers are akin to the AMC Pacer and Gremlin, and this parallel can't be lost on Sun's only non-tech founder. If you want to know what has driven Scott McNealy all these years, the ghost of AMC is at least part of it.
As my wife says: Foreshadowing is how you tell the quality literature from the cheap stuff.
Let's say that this IBM acquisition of Sun does come to pass. What on earth does it mean for the future of systems?
It is very hard to say. Obviously, controlling Java, Solaris, and MySQL, the three key open source programs that Sun has under its control and that it has monetized somewhat is an interesting thing, and IBM could so some interesting things with OpenOffice as well if it wanted to take a run at Microsoft's desktop empire. There is plenty of good engineering in Sun's Sparc and x64 server lines, and the revenue stream of support contracts for Solaris machinery is nothing to shake a stick at (although it pales next to IBM's much larger revenue stream for the wickedly smaller number of mainframe footprints in the world). Solaris, I think, is the key to Sun's empire, and the fact that the company forgot that many times in its own history just goes to show you what happens when you don't remember who you are and why people give you money.
I can envision IBM putting Solaris on Power Systems machines as well as mainframes, but there are application porting issues here. Even if you use the QuickTransit emulation software Big Blue picked up last fall to run Solaris applications in emulated mode on Power or mainframe iron, this is not the same thing as native. So IBM could, oddly enough, keep the Sparc platform alive--or at least not kill it. I think it far more likely that IBM keep all the Sparc intellectual property, license it to Fujitsu, and have Fujitsu make the Sparc chips and servers that IBM then rebrands and sells--much as Sun does today for Sparc64 VII-based machines. A lot depends on how good the 16-core "Rock" UltraSparc-RK processors and their "Supernova" servers really are. They are due to ship in the second half of this year, and present the first credible threat to IBM's Power Systems in about nine years.
What I can't see is why IBM would pay something close to $8 billion to kill off Sun. IBM did, you will remember, pay $810 million to buy Sequent Computer Systems, a company that had its own Unix server line as well as a nifty Windows server business that is the foundation of the high-end System x business (only loosely) today and which, I believe, also contributed technology to the Power 570 and earlier machines that bear a strong resemblance to Sequent's NUMA-architected iron. IBM ripped the guts out of Sequent and killed it, unlike its acquisition of database maker Informix, which sort of lives on inside IBM's Information Management division within Software Group.
I think this deal is a bad idea, mostly because I like competition and strong competitors with different ideas about how computing works. But that may just be an idea whose time is passed, thanks to the economy. The ascendancy of companies like Sun and the bad economy almost killed IBM in the early 1990s, and it may just end up killing Sun, a much smaller company than IBM was back then, this time around.
No one knows exactly what is happening right now between IBM and Sun. The latest word from the Wall Street Journal is that IBM's lawyers are pouring over all of Sun's contracts and licensing agreements to see the impact of an acquisition and what they actually get if they decide to buy the company.
The Four Hundred is on hiatus next week, March 30, but if Sun does get eaten by IBM, then I will post a story in the Breaking News section on our home page telling you what is what.
Post this story to del.icio.us
Post this story to Digg
Post this story to Slashdot