Will Infrastructure Outsourcing cease to exist?

Writing in Horses for Sources, Thomas Mendel writes "Cloud is already mainstream and legacy infrastructure outsourcing will cease to exist"  (horsesforsources.com/winners-circle-cloud-infra-svs_050814). Now, Phil Ferst et al have made a living out of hyperbolic proclaimations, but this one really caught my attention: to what degree is it accurate?

While Network Services is the fastest growing segment of IT, Data-related real estate (previous known as "Data Center") has to be the most transformational. For example, SunGard, that perennially financially troubled DRP player, has a shrinking market even though data is exploding. Why? By vastly syndicating data assets, risk and vulnerability has shrunk. Compared to the days when tape backup and storage was relatively expensive, enterprises can now protect their data by managing the sources and residence.

The article is interesting because of who is in the magic-quadrant lookalike (okay, it's pretty useful and I'm jealous that he came up with a credible alternative to the Gartner franchise); and who is not in the "Magic Circle."

Based on their ability to integrate pure cloud; hybridge 'hosted;' and private facilities, the top tier are AWS, Cisco, HP, Accenture, Dell, Wipro, NTT, IBM, and a couple others. AWS and Cisco are new entrants over the last few years, how is that they bolted ahead of the next tier, primarily dominated by telcos (ATT, VZ, BT, OBS)?

And what does it mean for Sourcing Advisory? Contracting for Cloud services is not your grand daddy's sourcing agreement. Cloud services tend to be fairly rigid - suppliers keep things consistent. So the legacy advisors expecting to build highly customized SLAs and provisions will have a long, hard slog trying to negotiate for their clients.