IBM in the past has seemed more interested in maximizing the value of their current technologies than trying to predict and guide future technologies.

For example:

1. They missed the move to Cloud and eventually shifted focus to "hybrid" cloud, which allowed them to keep selling their existed software and hardware.

2. They missed the move to SaaS, trying instead to sell expensive on-prem enterprise software like Notes and Commerce, both of which they eventually sold to HCL.

3. Even though they were pioneers in virtualization, they were more interested in bundling their software on expensive appliances than enabling it to run in VMs.

The one thing they can predict is when something is on the verge of commoditization, at which point they decide it's too difficult to make money on and sell it off, e.g. PCs, laptops, hard drives, POS systems.