1. Fusion
2. Fission byproducts
3. Catching the solar wind
4. Mining the atmospheres of gas giants
>1. Fusion
I appreciate your optimism but unless you're talking about harvesting the fallout of fission bombs fusion probably won't be a practical source of helium for quite some time.
As for harvesting gas giants imagine the costs involved to send a ship close enough to one of these monsters to harvest a significant amount of helium, then manage to escape the gravity well and make it back to us. That would probably be end up being the most expensive substance on earth.
Of course that’s currently only worth ~6,000$ so it’s not being collected. But, if that’s the only option at 100,000x current prices it might be used.
Hypothetically, if 30% of the world’s electricity was generated by fusion they would be producing ~350,000 kg of helium per year. Still well below current production, but also not an insignificant amount.
A more efficient design would probably be to have a miner full-time there, sending payloads in higher orbit, and a transport ship that approaches on a parabolic orbit, so it only needs engines to accelerate the payload, counteract friction and adjust trajectory
It would still be very expensive, though.
I suppose that the good news is that wherever you're getting your helium from you'll probably also have copious quantities of hydrogen available so you won't have to carry that over. Managing to oxidize it however is left as an exercise to the reader (apparently there's a tiny amount of water in Jupiter, but of course then you'd need some other energy source to electrolize it).
Or you just push it through a pipe, space-elevator style, exactly how many hard science fiction authors do it for mining gas giants for reaction mass/other fuel.
Should be considerably cheaper from an energy standpoint.
If you're to a point where you can create such a station and manage regular transport between two bodies, you're going to be able to build some sort of pipe that should be able to contain the helium enough to pump it far enough out to considerably reduce the cost of leaving orbit.
For a visual representation of this, see https://xkcd.com/681/
So, one could imagine a siphon which would take helium from Uranus to Earth. A normal siphon moves a liquid out of a gravity well by relying on a deeper gravity well. Of course, there a limits to the height that siphons can handle, based upon the vapor pressure of the fluid, and the idea of running a really long tube from one planet to another is kind of ridiculous. However, the same energy differences could be used in principle to power a more realistic scale system, like a recirculating ship which gains kinetic energy on the Uranus-Earth run and loses it on the pick-up-helium-at-Uranus run. Such a system would need to dump angular momentum around the Sun at one of the planets, because this doesn't balance when you move mass from one planet to another.
My inner five-year-old loves the idea of mining gas from Uranus. :)
If you managed to produce ten thousand tons per year (out of a market of about 25,000-30,000 tons per year currently, although perhaps set to grow along with the global economy), the cost could be manageable.
It depends on launch costs and cost per kg of the material. Market price of refined helium is what, around $40-100/kg? Maybe up to $400/kg in the case of supply crunches and in refined liquid form?
And it probably depends on development of nuclear thermal rockets/ramjets to enable you to reach orbital velocity of Uranus in one or two reusable stages. As well as improved thermal protection systems (TPS) to enable reusable TPS at the higher velocities of reentry for Uranus. Currently, with chemical rockets, we can probably get costs down to $10/kg for bulk liquids to LEO (see here: https://www.spacex.com/sites/spacex/files/making_life_multip... slide 41, do the math and the tanker costs per kg of propellant to LEO is less than $10/kg to LEO).
It's not impossible for nuclear thermal rockets to get prices down low enough. Even assuming fairly low efficiency (say, 1000s Isp with about 20 kg of propellant needing to be expelled for each kg of payload to escape velocity... taking into account dry mass ratio of about 20 due to the exponential rocket equation and Uranus's 21.5km/s escape velocity) and non-breeder reactors (where the fuel cost is about $0.15/MWh), the energy cost of the uranium to push a kilogram of material from the cloud layer of Uranus to escape velocity is still on the order of cents per kg. Of course, it then requires years in transit, so it'd only be feasible if financial/market conditions were exceedingly stable.
So in principle, it's possible to imagine that in 100 years hence, it could be feasible to mine Uranus for helium and meet near-current market prices. But the current ~$2 billion/year market for helium is probably much too small. But when it's a more-lucrative $20 billion/year market? More realistic.
But helium-3 for low(ish) neutron fusion may be a larger market by then, and THAT could be a significant market as well, with less challenge for getting price per kg to orbit reasonable (say, on the order of $10,000/kg or more for energy purposes, and it may be feasible to use expendable rockets for such small amounts of material, thus avoiding the more extreme performance requirements for the TPS and structure). Perhaps helium-4 mining would be an off-shoot of helium-3 mining, then.
But you forgot a few: mining asteroids, catching comets, processing huge amounts of lunar regolith and prospecting for Earth sources. Which are almost possible and economical, compared to the rest. In fact the latter is being done.
none really