In this post, Jay Manifold comments:

…my (possibly incorrect) understanding is that LH2 can only be stored for a few hours.

It is incorrect. There’s no intrinsic limit on how long you can store LH2. It’s just a matter of how much weight and power you’re willing to devote to insulation and/or refrigeration systems.

In fact, the concept I have for a cis-lunar infrastructure is a series of combination depots/tankers. You’d have at least four of them (probably five, for backup purposes). One would be sitting in LEO, being filled up. One would be sitting at L1 to provide propellant for returning and lunar-bound vehicles. There would always be (at least) two in transit, one heading toward LEO, and the other heading toward L1. When one arrived, the one already there would depart to the other destination.

I’m envisioning them with plenty of power, both to run high-Isp thrusters, and to keep propellants continuously chilled. You might be able to do it with solar (though the panels would take repeated beatings going through the Van Allen belts). The obvious technical solution would be nuclear, but that’s probably still politically unacceptable, despite its reasonableness.

[One further evening thought]

If they were powered with nukes, there’d be plenty of power to not only keep the hydrogen chilled, but to actually crack it from water. The marginal cost of doing so, given the initial investment of the nuclear tankers, would be pretty low, and it could dramatically affect the cost of delivering the propellants to orbit, since they could be delivered in a more dense form that doesn’t require cryogenic tankage, and is much safer. Of course, the vehicles would either have to operate at a stoichiometric ratio of 8:1 oxygen/hydrogen (which is suboptimal in terms of specific impulse–ideal is 6:1)) or throw away or find other uses for the excess O2.