Several things:
1) There's a general consensus that lots of workloads will never go into the public cloud. Whether or not Openstack becomes the de facto standard for in-datacentre computing (and I'd think VMWare would be one of many who'd have something to say about that idea) the idea that public cloud will just sweep away the entire ecosystem of enterprise vendors and resellers away is absurd, however much headway it's making with the low hanging fruit.
2) NASA's leadership in Openstack was actually a key factor in their leaving. Suddenly all of the software engineers who'd had anything to do with it were headhunted for big salary increases or left to start up companies. A huge disruption and cost increase in a non-core area of operation would cause any enterprise to look for off-the-shelf alternatives.
Whatever the public vs private economics for NASA given the datasets it works with, the question of whether it had any business continuing to lead in a market that isn't core is different from the question of whether more mature and vendor supported private cloud options are economical now.
3) I've never seen anyone outside the press talk about Google's cloud. I'm sure people are using the free credits thrown at them to try things out, but is it really gaining any traction?
Similarly, I don't see Microsoft as wholly hostile to the wider market. Azure is definitely getting traction, but the overwhelming majority of Microsoft's revenue still relies on the people this article places in the opposite corner.
That means AWS is really in the driving seat on public cloud and it's a formidable competitor, but I predict that the more it tries to expand to attract traditional enterprise computing workloads the harder it will find the going. Private datacentres are here to stay for well over half of the compute market (at the very least) and lots of the software as a service apps which spring up to fill gaps will grow to a scale where public cloud isn't a fit for them either (see Salesforce, for instance).
The really interesting thing with AWS is not really the size it's grown to (though it's impressive) but the sheer scale of mindshare and disruption that it's acheived despite being far from the monopoly position that Microsoft acheived on the desktop.