In the end, Cloud is just IBM Service Bureau, without the leased lines (and the security of same). The only way for Cloud providers to make any money is to shift cheap goods sold dear: the lowest common denominator of each resource and top 1% prices. It may work for a while; while clueless CIOs chasing the latest Intelligent Business meme have a job.
For Cloud, as hyped, to work, the workload has to be steady state. And that means that spikes for resources have to be balanced by dips for those same resources at the same time (for some time delta). The odds of that actually happening, in general, are slim. If it were a slam dunk, The Service Bureau would still be around raking in cash. It isn't, and it isn't. In the end, CIOs will have outsourced some capex for opex, and lost control of their function. For short sighted (and termed) CEOs, the quarterly report might look better, but in the end, the IT resources needed to do the job are fixed by the nature of "the job". Outsourcing just adds the outsourcer's profit into the cost. For bureaucracies, that means a larger budget (in opex), which means more "power" in the organization. It's xml schemas, only lots worse: a bad idea executed badly.