Earlier this week, I speculated that Oracle's statements that it was keeping Sun hardware would eventually lead to the company needing to buy or merge with a desktop hardware vendor. I concluded that if that were the case, Wyse would be the likely target. However, this week I had dinner with several ex-senior Sun executives and they said that what I didn't realize was what Oracle's goals were and that what Oracle was saying was intentionally misleading. I found their arguments compelling and think they should be shared.
As I pointed out, this move into hardware for Oracle was incredibly risky. Hardware margins aren't what software margins are and would adversely affect Oracle's valuation. In addition, hardware and software are vastly different. Oracle, as an institution, has little real hardware experience. Its experience with thin clients showcased this and represents one of Larry Ellison's greatest failures. Oracle's strength is software, as its most recent acquisition of Virtual Iron points out. And as I also pointed out, the problem with hardware is that it turns partners like Dell and HP against Oracle and would probably benefit Microsoft, which would be used by both to offset the increased risk that Oracle with hardware would represent. Oracle knows this and the math suggests that it would be vastly better off dumping the hardware. So, why would it say it wants to keep it?
Something to Sell
The ex-Sun executives argued that the remaining Sun hardware business was probably how Oracle justified paying more than IBM did. And this is because IBM could only shut it down -- but Oracle could resell it once it took control. That business is likely worth $1 billion to someone, particularly if it came with a tight relationship with Oracle. But with every lost Sun hardware customer, this value drops. And without the perception that this business would continue, this asset would bleed customers like water, leaving an asset worth a fraction of what it now is.
Since Oracle can't sell something it doesn't yet own, talking about the eventual purchaser is problematic. The eventual buyer can't be effectively engaged until the merger closes. Even then, a lengthy process starts, suggesting that the sale would still be at least three months, and possibly in excess of six months, after Oracle takes ownership. By then, Dell, IBM, and HP, among others should have divided up the majority of Sun's customers, leaving something that would be hard to sell, even for pennies on the dollar.
Therefore, Oracle has to maintain this business until it can sell it and can sincerely say it has no intention of closing it down. But this means that it has no intention of keeping it either. It simply wants to maintain as much value as it can so it can maximize the eventual sale of the unit. It is likely Oracle has explained this strategy to companies like Dell and HP to keep them from moving competitively against it and to open negotiations with them to sell the property. Both companies would rather the other didn't get this asset if it is worth anything significant, and it wouldn't surprise me to find out that Oracle was playing vendors against each other to maximize the selling price. These executives I spoke with did think that Oracle might want to retain the Sun storage business if it retained anything, which is an interesting final thought.
Why IBM Didn't Buy
One part of the discussion I didn't entirely agree with was why IBM wanted to reduce its purchase price. The ex-Sun executives indicated that in the months leading up to the announcement, employment contracts were altered so that senior executives got millions out of any sale. They argued that IBM HR likely had a cow when they saw this and that the lowered bid reflected the reduced value associated with the cash payments to executives and to make sure IBM didn't appear to doing anything questionable. Larry is vastly more flexible with payments like this; he bought PeopleSoft just to shut it down, after all, and likely felt a few million dollars to executives were more of an ethical problem for the Sun executives and the Sun investors then they were for him. Given the stockholder lawsuit, I wonder if that was an accurate assessment.
Keeping Sun hardware really didn't make sense to me. Oracle knows better. While it was fun to speculate about it recreating another IBM, even IBM isn't recreating the old IBM. It is more software and services today and vastly more profitable as a result. Larry isn't stupid and it simply seems more likely that he will sell off the hardware unit than keep it. Why IBM didn't buy is interesting but largely academic. However, the litigation could delay the acquisition enough to destroy the extra value that Oracle thought it was buying, making IBM look smarter in hindsight. It goes without saying that you should probably watch this all from a distance. There is simply too much risk until everything settles.