I've been working in the private space business for years, and one thing we figured out early is that you can use technology to drop transportation costs at least as fast, and usually faster, as you can use it to develop on-site materials usage on the Moon or elsewhere.

After a long time, that starts to change. On-site solar power, of course, is always more practical in space. On the Moon, on-site construction materials for things like buried bunkers and such start to look good fairly quickly. Water ice doesn't suck either for things you need water for, or for LOCAL fuel use (cracked by solar power). Cracking oxygen from the regolith is way too much work, especially if there's ice.

Helium 3 has always been a joke for near-term use. We don't even have working fusion reactors for the simpler hydrogen reactions, and Helium 3 is even harder to fuse. It's like having an oil well in the year 1700.

A lunar catapult system changes the economics a bit more, making things like ice for fuel elsewhere look better.

Nobody's going to be making aluminum cans or microchips on the Moon for a very long time. Actually, some asteroid metal resources may be more practical to exploit sooner than those on the Moon.

But who knows what industries might eventually be found to be perfect for the Moon? Nobody would have predicted Las Vegas seeing the Nevada desert 200 years ago, or that the Pacific Northwest would one day be a center for smelting aluminum.

For Mars it's a bit different, since the transportation hump is a lot higher (barring things like nuclear propulsion), and I think oxygen and water will be easier to come by.

For both worlds, hydroponics and other methods of using plants for food and air won't be a bad investment, but you would NOT want to rely on them 100 percent for either.