What took them so long?
When the (scientific software) company I used to work for got onto the Cloud, we had a heck of a hard time getting our major pharmaceutical customers to buy into it. They were concerned about security. We would ask, "Do you really think your own security infrastructure is more robust than Amazon's?" There were some lead adopters, but not many, and yet many of these same companies were outsourcing their internal IT to outside organizations, especially IBM, in what was sometimes called an "insourcing" arrangement.
We had close relations with IBM at the time, had historically supported AIX, etc., and I remember telling our contacts at IBM that if you had a cloud, all your existing IT-management customers would buy in, because they would be willing to trust you, even if they don't trust Amazon. IBM would respond with some nonsense like "Well, we already have a cloud." It was not by any means what a cloud had already come to mean at that time. What they had was a farm of supercomputers, fixed partitions of which you could lease by the day, week or month, based on what might be a long reservation list. A cloud implied leasing virtual equipment by the hour, the provision expanding and contracting per demand.
Their Cloud-Pak and OpenShift technologies sound good and are somewhat orthogonal to IBM's public cloud, but I feel they missed a big opportunity not that many years ago.
Admittedly, it takes a lot of water to turn a big ship around, plus they were frying other fish.