Mad Dog 21/21: The Age of Acquire Us
May 4, 2009 Hesh Wiener
It looks like Oracle is going to buy Sun Microsystems. This will probably happen soon, but it’s hard to be sure. Not long ago it looked like IBM was going to buy Sun, at least according to a passel of ace reporters, but it didn’t happen. However as events unfold, one thing seems clear: Sun’s role in computing is huge. For an America suffering economic collapse, Sun is a necessity, which makes it a mother of invention. Sun is akin to hope, that strange invention, which may have feathers. And we need hope so much right now.
Until now, Sun has been a buyer rather than a seller. It acquired MySQL in 2008. It took over StorageTek in 2005. It bought OpenOffice (which at first was called Star Office) in 1999 and released an open version of the source code the following year. It even talked about buying Apple once, back in 1996, but did so in an unrealistic tone.
Now the tide has turned on Sun as on so many other companies in technology and elsewhere, in the USA and outside it, too. Share prices are down, making good businesses with diminished market value into acquisition targets. Moreover, companies that are no longer as confident in the future as they formerly were don’t necessarily mind a little attention from prospective buyers.
Sun is up for sale because it is scared, or at least its directors are. The company’s board can’t turn down offers that would just about double the value of its shareholders’ equity, not without a pretty good story and lawyers to match. The possibility that Sun would be bought by IBM for a lot of cash and, in the rosiest scenario, become a thriving part of an empire that says it will weather this economic storm, really caught the attention of the investment community.
The IBM deal might not have gone over as well with Sun’s customers, who presumably had good reasons to choose Sun over IBM as a vendor. Sun’s users might have been uncomfortable if $14 billion JAVA (its trading symbol) was replaced by $103 billion IBM. The mainstream media seemed to presume that IBM would preserve Sun’s hardware, keep pushing its impressive portfolio of software, resolve corporate cultural conflicts, and that everyone would live happily ever after. After mulling it over, though, IBM said no dice. Some of Sun’s shareholders may have wept, but we have to figure that plenty of Sun employees and most of the company’s customers were relieved.
A decade ago, IBM bought Sequent, an inventive company that made fault tolerant computers based on Intel chips. IBM took advantage of the technology behind Sequent’s NUMA memory architecture. The rest of the company became a dried husk, like something eaten by a spider. Sequent had a relationship with a company called Fundamental Software that had written some pretty good mainframe emulation code. That code, Flex-ES, ran on Intel boxes running SCO Unix and later (for developers, anyway) running Linux, too. But factions within IBM always resented Flex and felt that IBM should never have endorsed any emulated mainframe. This is so even if some academic types consider all IBM mainframes from the System/360 onward emulations of a sort because they use firmware to create the mainframe surface that obscures various underlying processors.
More recently, IBM bought Platform Solutions, which emulated IBM mainframes using firmware, software, and other magic made possible by Itanium chips; Platform sank with barely a trace into a Big Blue sea.
Sparc chips and other Sun inventions would be harder to ditch because Sun has put a lot of its key CPU technologies into the public domain. Sun is also big on open software, and has set things up in ways that permit its own contributions, such as Java, and the contributions of open (or subsequently opened) firms it acquired to survive no matter what the future holds for Sun. That may be part of the reason IBM’s sage outside counsel, the law firm Cravath, Swain & Moore, seemed to be hovering over IBM’s executives when they decided they weren’t interested in their brutish empire setting on Sun after all.
IBM also may have had concerns about some of the antitrust issues raised by the possibility of a Sun acquisition. Big Blue’s conduct is the subject of a European complaint brought by Florida-based T3 Technologies, a small company crushed by IBM during its recent anti-emulator campaign. Here, again, Cravath’s caution might have played a role in IBM’s decision to abandon the chase.
In addition, Sun bought StorageTek in mid-2005, making Sun a significant rival in enterprise tape. Ironically, IBM probably would be just as happy if somebody, including even Oracle or EMC, took over its own tape business, a segment in which IBM probably finds maintenance is the most profitable corner. The tape trade is under a lot of pressure these days, because efficient X86 servers can economically do the job that IBM tape controllers do with more costly Power-based servers. As the X86 iron takes over, IBM won’t be able to compete on a cost basis with Power hardware; its only hope will lie in software that is tied to its more costly Power-based subsystems, and it’s not clear how long IBM can hold a lead by that means.
Sun has some excellent processor designs that include its eight-core, 64-thread “Niagara” chips. Sun is also the shepherd of the platform-independent Java language it developed. IBM seems to be a big fan of Java and supports the language, as it does the Linux operating system, as part of its posture of openness.
Now that IBM has merged the System i and System p into a single Power Systems line and started building mainframes whose processes share similar components (but which have different cores and instruction sets), it’s pretty clear to anyone who cares to look that IBM has a lot more hardware revenue from Power than Sun does from Sparc. Sun customers had good reason to fear that, had the acquisition gone through, IBM would have let Sparc fade away, at least in the IBM product line, even if Sparc technology persisted elsewhere.
Sun and IBM both sell X86 and X64 type servers. IBM is more of a household word in glass house country, but Sun’s Intel and AMD servers get a lot of respect from users who have full heads of un-gray hair. But the big question is not whether Sun’s ideas would help IBM enrich its X86 and X64 Series x line but rather whether IBM would just, after a decent interval, end up selling off the whole shebang to Lenovo or some other particularly efficient manufacturer.
The Sun customer base and IBM customer base have big cultural differences. These differences exist even in companies that use IBM in the back room where data processing gets done and Sun in the front of the shop where interactive computing, including Internet functionality, is job number one. Sun’s customers are smarter, or think they are; IBM’s customers are richer, or think they are. What happens when you add Oracle into the mix is anyone’s guess. Oracle’s customers are both smart and rich, and they are much more attuned to the interactive Unix and Linux worlds than IBM’s customers, even though IBM has spent a decade cultivating the Linux community and its Internet relatives, such as Apache.
It’s very hard to precisely predict how Oracle is going to deal with MySQL after it consummates the Sun deal. Oracle’s inclination might be to use MsSQL to drive Microsoft nuts by enriching it a bit and then positioning it directly against SQL Server. Microsoft is working hard to get its DBMS systems into glass house settings, seeking out situations where IBM’s DB2 is viewed as a luxury. To do this, Microsoft is enhancing SQL Server so it can take full advantage of emerging midrange and large multi-threading servers; otherwise, IBM can make the case that even if DB2 costs more, it is a superior product. Oracle is ahead of Microsoft when it comes to rich functionality and support for large servers, but Microsoft wants to close that gap, too. MySQL is not the high-end answer. In most cases, MySQL cannot use all the engines and available threads on a single Sun Niagara chip or a high-end X64 box; it is an even worse match for servers that use groups of most elaborate Power chips. That might protect Oracle but it also gives Microsoft openings Oracle might be wise to block.
IBM has from time to time paid attention to the database business in midrange and low-end markets, but lately it has slipped, letting the i-on-Power fall behind Windows-plus-SQL Server (and Linux or Windows-plus-MySQL) alternatives. It’s only going to get tougher in that part of IBM’s market now that Cisco Systems says it will be moving into the server trade, too. If IBM decides to really fight hard in the midrange to preserve its Power line (instead of concentrating on high -end Power systems and mainframes), then its decision to abandon an acquisition of Sun starts to make more sense. IBM needs more focus, not less; Sun would offer IBM as many diversions as opportunities.
Another factor Big Blue might have considered when it got serious about buying Sun: IBM might not be able to afford the acquisition. IBM already has more goodwill (about $18 billion) on its balance sheet than shareholders’ equity (about $13.7 billion). Buying Sun could have dented IBM’s balance sheet by boosting goodwill by another $5 billion to $23 billion, making goodwill about twice IBM’s net worth. Investors might not tolerate that. IBM is a very strong company. but these days even IBM can’t really afford to ask investors to ignore the past year’s erosion of its balance sheet.
Perhaps surprisingly, Oracle can probably handle the financial impact of a Sun acquisition better than IBM. Its net worth is $23 billion, nearly double IBM’s. It has a bit less foam on its books; goodwill is about $17 billion. And it has more flexibility than IBM when it comes to adjusting its pricing and product lineup because it does not have a huge services business. IBM’s services operations are a great source of what IBM calls annuity revenue, but they also anchor the company in ways that keep it from moving at full speed into disruptive markets, such as cloud computing. IBM has to consider its thriving but backward-looking services business whenever it seeks ways to compete with Amazon, eBay, Google, Apple, and other companies that don’t have a huge vested interest in the legacy bookkeeping systems that IBM manages through its services operations.