I strongly suspect the chances of Orion-style propulsion happening are close to zero. I think we're much more likely to see fission engines (where the heat from a conventional fission reactor is used to superheat water, which then goes through a conventional nozzle), or something a little more exotic like pulsed fusion (where little capsules of lithium containing deuterium go through a magnetic pinch and fuse, again in a more-or-less conventional nozzle) are more likely. They also have very high ISP but with much lower abuse risks (can you see any government agreeing to have a giant stockpile of nukes in space?)

You are probably correct, but it still remains that fuel exists in stockpiles and it isn't a particularly complicated spaceship design for the phenomenal payload and "extremely high performance".

One starship trip could probably transport enough "fuel" for a midsize orion trip to mars and back (I haven't run the numbers, but 150 tons of fissile nuclear fuel is pretty effing compact.

The "Super Orion" 8 million ton ship uses 3,000ton bombs.... but most of the bomb is the inert propellant that can probably be obtain from regolith.

And the Moon could probably have a nuclear launch cannon pretty easily to reduce the payload or initial acceleration.

Politically it sounds dangerous putting nuke bombs on the moon, but frankly it is safer having them there than on Earth. You know, as long as a mass driver catapult doesn't send them back ....

Oh, but to your point, one of my ideas for Starship is to design an antimatter catcher from the solar wind:

https://www.niac.usra.edu/files/studies/final_report/1107Jac...

So you could probably do the antimatter catalyzed drives a lot more feasibly in space, but its still a lot more complicated than dumb bombs on pusher plates.