Why IBM might ditch servers and become Intelligent Business Middleware

The tides of enterprise IT have shifted. While there’s still money to be made in selling servers, the innovation is now happening higher up the stack. Recognizing this, it looks like IBM has decided not to fight for share in a shrinking or stagnant market, but to get out and focus on the future.

Analysts and reporters are speculating that IBM is in talks to sell its server division to Lenovo, the company that IBM sold its PC division to back in 2005. Lenovo has confirmed it is in talks with an unidentified buyer in a filing to its shareholders. From a Wells Fargo analyst note published Friday morning:

We believe an acquisition could give Lenovo greater scale in similar components to its PC business, which could provide greater margin or the ability to pass along the benefits to customers via lower pricing. This, in turn, in our opinion, may cause further competitive pressures in the industry and have adverse impacts on Dell and HPQ. For IBM, it would be in keeping with its historical strategy of divesting commodity-like businesses. While we would fully anticipate some purchase commitments by IBM, this may signal a longer-term trend of moving away from x86 to a greater focus on System P for its solutions-based products.

Getting out of the server business as it moves closer to becoming a commodity — with fewer customers who are unwilling to pad a hardware maker’s margins — makes sense. While Dell tries to go private, and HP doubles down on the enterprise with servers such as its Project Moonshot offering (which isn’t anything a webscale vendor would want), IBM is looking ahead.

WatsonAnd what is Big Blue eying? It’s already told us for the most part: big data and mobility.

IBM has spent billions on analytics. It’s researching ways to push the boundaries of memory chips and new types of processing more appropriate for real-time information flows. It’s even building intelligent systems such as Watson. It also has been buying companies such as WorkLight and CastIron Systems to help tie cloud-based applications together and ensure acceptability in a world of federated apps.

IBM is building two things here. One is a near-term strategy that can be thought of as next-generation middleware that will tie mobile applications, the underlying mobile (and wireline) networks and the cloud together in ways that will give enterprise customers the sense of accountability they need. Things like ensuring that APIs are dependable enough to build service level agreements around and secured if needed isn’t sexy, but IBM is taking it on.

Over the longer term, the company is pursing efforts like IBM’s new memory chip research and new processors that work more like the human brain. Watson, which will be delivered eventually as a cloud-based mobile app and likely rely on some of the middleware related knowledge, is the first of IBM’s longer-term plans. Manoj Saxena, the head of IBM’s Watson division, calls this the fourth wave of computing.

And while there will be servers aplenty in that vision, they won’t look like the boxes of today, and they certainly won’t be the profit-centers. The value will be in the services themselves, the hard work of integration and perhaps the new silicon (or graphene or DNA-based) platform on which these things all rest. Today it’s middleware and tomorrow it will be a revamp of the machine.