But what if that move is a strategy that is cloaked in brilliance? Hear me out on this; it might be that IBM is just a little smarter than some people give them credit for.
Disk storage--or DASD, as we like to call it--is profitable to IBM on a few platforms, iSeries being one of the most profitable. The disk drives are essentially the same on the iSeries and the pSeries (formerly RS/6000). The volume of drives manufactured are used in the PC marketplace. Dell, ThinkPad, Aptiva, and others use IBM disk storage in their PC clones.
So why can't IBM make disk storage profitable? I think they can. I think that the disk storage unit is profitable and that it is making big money. But the problem may be that IBM believes that electromagnetic technology is going to reach its threshold within five years. If 18 GB disk drives from IBM for iSeries cost $4,000 today, what happens when they have 500 GB disk drives on the PC or ThinkPad that cost under $1,000? What happens if at 1,000 GB they run into a wall in the electromagnetic technology?
Theoretically, let's say that 1,000 GB disk storage is it. Suppose they don't know if they can make denser disk drives using current technologies. Yet, in the disk storage business, you have to keep increasing the density. You have to keep coming out with denser drives and charge a premium for them to offset the deep discounting at the low end. But what if you can't? Do your profits drop as your prices drop? Definitely.
So what is really going on? Well, I don't know, but I think it may have something to do with IBM's Millipede technology. This nano-technology is an entirely new way to store data safely. IBM has already demonstrated that by using Millipede technology, it can store 20 times more data in the same space as it can on most advanced disk storage devices available today.
Millipede storage devices are not as fast as today's high-speed DASD, but they do seem to provide the ability to support the equivalent of multiple "arms" per "platter"--meaning that while a single disk drive may be faster on a one-to-one comparison, by adding additional read "tips" to the Millipede device you effectively cut the data access time. Adding two additional read "tips" means even better access time. Imagine a storage device with dozens of hundreds of read/write "tips." Also, Millipede storage devices don't seem to be impacted by electromagnetic influences. This means that you could place a magnet on one of the storage devices and it would have no impact on the data stored there, an important feature for military applications.
So is IBM off base in selling its disk drive manufacturing unit? Or is this something uncharacteristically clever on IBM's part? After all, how much would an old disk drive manufacturing facility be worth if Millipede technology does become commercially available in two short years?
I think IBM may be smarter than we think.