> Keeping multiple companies capable of building it alive is essential.
Companies and the capability of building are two separate things. It is not at all a desirable thing to keep a company alive which refuses to develop and implement the capabilities to compete, in the process depriving resources from those that would develop those capabilities. If a company dies, its talent and equipment do not vanish into thin air, they get bought up by competitors who can put them to better use.
Unless you are actually duplicating efforts to have multiple firms produce the same things, a large number of potential suppliers does nothing to reduce your risk once you select one to move forward - especially if you still are required to use them after repeated failure. There are just a greater number of potential failure points as any of your suppliers, all of whom you rely on, might fail.
Further, in spreading contracts out among many firms, you reduce the economies of scale of any individual firm. They can not build out the additional capability that more work would afford them, all the while they are taking resources away from genuine productive capability by duplicating effort with excessive overhead.
Concentrated monopolies are bad for common consumers, who have no negotiating power and can be extorted. Governments don't have that weakness. On a purely economic level, the government is a single buyer - it's a heck of a lot easier for them to find a new rocket maker than it is for a rocket maker to find a new government that will buy from them. Beyond that, governments have a monopoly on violence, piss them off enough and bankruptcy is the least of your worries. If it really wanted to, the government could just do the work in house, either setting up new public firms or nationalizing existing ones. Excessively costly government contracts are graft, or at best pork; the government could easily get much more favorable terms if its leaders were so inclined.