The primary lift would be from the rocket engines. The use of gaseous H is merely an attempt to turn a hazard into a useful asset (a lighter net launch weight.) It only has merit for this unique high volume and relatively low weight rocket/payload design.
For more dense payloads that can stand high acceleration certainly the various mass driver concepts are more efficient over the long haul. Different solutions for different payload requirements and constraints. Both extremes of payloads will be needed to launch different things needed in space. The launch of dense commodity items is relatively straightforward. It is a much more difficult task to bring more delicate large scale items into space, just as it is transporting similar items terrestrially.
And of course the economic transport of people and other high value, moderate sized, delicate items to orbit is still the missing piece of the launch line up. It still costs way too much to get us up into space. That is the critical challenge that needs to be overcome.
I’m certainly not the space expert you are, that’s for sure. As for your venting the excess H into a blimp instead of wasting it(or recapture w/the cap), that may be a useful idea. Since the rocket would be relatively short compared to the blimp, perhaps the pumpkin balloon configuration may be better suited to your design. Even if it only gets you to 30,000 - 40,000 ft, that’s still a lot of fuel saved.
Besides a UFO there are basically 2 ways to get to LEO : 3 g on the shuttle, and EMSL projectiles at 100,000 g to 250,000 g; something like the passenger and freight cars on the train; different handling techniques. During the heyday of EMSL(during the Star Wars/SDI era)all the possible components were g-tested. Only a delicate clock mechanism was damaged(that was expected).
My(hokey)design was a ring cannon about 1/2 mile in diameter at 15 deg to 45 deg tilt, built like a bicycle wheel lying on its side, straight Q tail w/about 15 movie films separating 1 psi step-up chambers(advance 1 frame per shot), thus maintaining internal vacuum. 1 g accel(13+ minutes to reach 5 mps injection velocity)w/machine gun burst(up to 20 projectiles/burst). Figured about 2 shuttle loads/day at 40 cents/#(10% efficiency).
It MIGHT have worked but Reagan’s SDI approach bankrupted the USSR, funding was slashed to near zero; and besides, NASA rocket scientists feared for their own space budget funding. Being shown up by EMSL as a better way to LEO gave them the willies. Thus they ignored it, and EMSL died in the cradle.
Still though, it was well researched by some brilliant people, I still have those 20 year old tech papers around here somewhere. The very best concept, in my opinion, was the quenched superconducting rings design. Basically it was a souped-up, modernized version of the old 105 howizter cannon, achieving 5 mps in a few hundred feet of length.
But it was all for nought, and nasa still makes $20,000/# to LEO for doing 4 KWH/# of actual WORK. Pretty good mark-up, yes?