If MS went under? Either someone would buy up the pieces, or someone like NSA would open source it.
I'm not totally familiar with the details, but I do know the NSA has had direct access to all OS source code cranked out by Microsoft. This isn't tin foil hattery, it's part of their core mission. They make technical recommendations to Microsoft to make the product more secure, so the US government has more secure computers. As not to play favorites, they have done the same thing for Cisco products, Oracle, etc. Hell, they wrote a variant of linux, SELinux that is a pretty good concept. If occassionally difficult to administrate. If you don't think that there's a vault somewhere on Ft Meade with the source code for every OS release, you must be taking pharmaceuticals on a recreational basis.
If Microsoft was somehow magically disappeared, the source code would probably be released or given to another company to maintain. For legitimate national security purposes. And enough businesses would lobby Congress to Do Something, that someone (or many someones) would take over. Maybe not well, but enough to limp by.
It'd be pretty much the same for any critical infrastructure across the country. If someone screwed up bad enough for the infrastructure company to bomb, someone would take over the job. May do a worse job, may do a better job. Probably have a few transfer problems, but they'd be sorted out eventually.
Remember Enron? Company exploded, but natural gas and electricity is still flowing.
So... Bill Gates then has no inherent right to go John Galt, and stick his closed-source designs in a fireplace when he's fed up with the world? He's too important? He has dominated the market?
(I know, some of you just threw up in your mouths a little at the parity of Galt to Gates.)
Isn't that a monopoly? Not only does it stifle competition during good economic times... it is detrimental to long term industrial and market safety during recessions or depressions. Here then is a strong case for open standards and separating your eggs out from just the one basket.
AIG was too big to fail? They had that much market share that they would have shattered the financial system had they gone under? The remainder of the insurance market could not have accepted those contracts?
I somehow doubt it.
And... if they were that big, then they had to have been a monopoly.
Let's stop chasing the Microsoft opensource/closesource argument and look at what a dominant vendor means to any market segment in the case of a collapse of that vendor.
AIG: Insurance. At any given time, how many contracts are actually claimed against? Less than 1%, I would hope. That means that the consumers of insurance simply need to stop paying AIG when the company sinks, and obtain a new contract elsewhere. No harm, no foul for 99% of AIG's customer base. Or whatever percentage are not actively pursuing a claim. Huge boom for the remaining insurance companies. Probably need to hire to accomodate all this new work.
Fannie/Freddie: Mortgages. A loan for X dollars, repayable over a typical 30 year span, for about 2*X dollars. In the event of default, reposession is very easy. Hard to hide or steal a house, and insurance covers destruction via fire or other causes. All in all, pretty safe market until you start lending to washouts in the school of life. Most of those mortgages, the vast majority, are in good standing. Let the company go under, declare bankruptcy and sell its assets (those loans) to level the books as well as possible. Borrowers pay Wells Fargo, BoA, et cetera. Whoever bought the paper on the house. Boom for the banks that didn't make idiotic lending practices a standard. How are they SO BIG that they can't be let go under? Deadbeat owners still default, whether the bank is solvent or not. Responsible borrowers still pay, whether the bank is solvent or not.
I fail to see any cogs falling off the wheels if these "critical" companies fail. And if they have more control over the market than their peers in industry then they can easily be muscling the market, thereby creating an antitrust condition.