Re: Wobbly Wheel
You can't make comparisons of wholly dissimilar things like that. That's how people who choose to self-manage their retirement funds end up bagging groceries at 79 years old. Incidentally, it's also why you can't compare say, Apple to Samsung beyond a couple of phones or RedHat and MS beyond that they both make operating systems. The businesses couldn't be more dissimilar really.
Your error is in not understanding what those P/L figures mean. Amazon isn't 'losing' money, they are investing in aggressive expansion. That is what keeps those revenue figure growing. Only a fool would stop that process before potential for growth had stopped. Amazon is incredibly capex heavy and investing now in operational economies that can be streamlined later is what will keep them in business after growth has plateaued. Profitability is a known point and can be attained in any quarter if that was required.
MS is, and always has been, exceptionally light on capex and high on margins. That's the upside of software. The downside is that if something goes really wrong those kinds of companies tend to collapse in big piles because there's no physical assets to move around/reconfigure (or even sell).
But both have the same goal, it's just they have two completely different roads to get to the money. One model has no superiority over the other. Both are extremely proved and validated methods, the person in charge is the key element, are they compatible with existing strategy?
Your last point is way, waaay off though. Managed infrastructure was never a suitable path for small entities. It's like Finland having nuclear weapons, insanely expensive, absolutely pointless and insanely risky with no reward. You only draw attention to yourself and pay mega dollars to be someone else's pawn, who they can squash in a second if you get uppity.
There simply aren't very many medium and larger businesses that actually make a good fit for hands off infrastructure. The risks are too great to lose control of your core infrastructure. There are very, very few of those businesses that actually have business models that actually include making money. Everybody is still waiting to see how that pans out. What I'm getting at, is that the only real money to be made in infrastructure services is going after small operations that have the talent to develop scalable offerings but lack the financial resources to maintain the requisite 3rd wheel of big IT that is support and physical plant (equipment).
Don't get me wrong, there's a lot of money to be made in addressing that market. Trouble is, that market, by default, can see only limited advantages from the economies of scale that are the sole advantage of offering scalable infrastructure services. The needs of the market are too widely varied for a big 'blaaargh' One-Size offering. You end up with a truly massive customer base where no two customers, even direct competitors, have the same needs.
The only way to crack that nut is to be huge. Just massive so that you can spread the inevitable losses around in the most effective, for the moment, way. The upfront and continuously growing investments are simply too large for anyone who doesn't already have the required funds and the intensely complex structure in place to play $1 Billion per hand 'go fish' with the stock traders and bank analysts.
Rackspace might end up being the infrastructure equivalent of those weird one-off mobile phone companies every big city seems to have. The ones in the dodgy parts of town where you go to by crack, assassins and prostitutes. Maybe the Radio Shack of consumer electronics retail. But they simply don't have the resources to compete.
Corporate finances are structured so that it's usually not readily apparent, but buried under the buzzword bingo metrics everybody likes to toss around, is their actual costs of making a dollar. It's normally a stunningly large figure if you can manage a peek behind the GAAP voodoo. If it costs you .96 to make $1.00 then your real estate agent probably plays golf with the guy who sold you your yacht and is married to the daughter of the helicopter conversion company that turned your Sea King into a hovering platform of GDP scale capitalism. That's petroleum or Apple money. For most successful businesses you're closer to .98 per $1. For mass scale utility infrastructure yore at .999998889+ per $1.
Funding just the accounting software to make such a thing possible is solely the domain of the stupendously wealthy. The kind of company you join early in your textile design career and 50 years later as a successful submarine designer you find out that every job you've ever had was with a company ultimately owned by the same people you designed Autumn season women's fabrics for 50 years earlier.
I guess that's a too long way to say there is not, and never has been room for small infrastructure providers who are more than flashes in the pan. It simply costs too much money to make money in that game.