Most fusion reactors are expected to use Deuterium/Tritium as fuel both because He3 is hard to source and more importantly because the D-T "cross section" is much larger than any other reaction: https://scipython.com/blog/nuclear-fusion-cross-sections/

See how D-T peaks at ~60keV? And the peak is much higher than any of the others, including D-He3? This means you need lower temperatures to achieve D-T fusion (keV is ~temperature) and larger cross section means that you're more likely to get a fusion instead of a scatter when a D meets a T.

D-He3 is promising because it produces no neutrons and so reduces requirements for neutron shielding and handling of radioactive material (but if you're fueling a reactor with D-He3 you can't stop D-D reactions from happening, and those produce neutrons, so you still need some shielding and contamination procedures).

> Sounds like China is planning to mine it from the moon.

They won't.

Sometimes you'll hear a quoted figure of "the moon has enough to power the USA for 1000 years". It might. But that's for the entire moon. To quote the linked article:

  This is roughly the same ratio of the isotopes as in lunar regolith, which contains 28 ppm helium-4 and 2.8 ppb helium-3 (which is at the lower end of actual sample measurements, which vary from about 1.4 to 15 ppb)

Fusion is about 1e7 times more energy dense than burning stuff, which makes this equivalent (in energy terms) to finding the moon's regolith made of (15*1e-9 * 1e7 = 15*1e-2 = 15%) 1.4-15% gasoline.

Extracting this from the regolith is highly energy intensive. It's so uncommon in the crust, that people seriously suggest using a scaled-up mass spectrometer; a byproduct of extracting it this way is pure oxygen and ridiculously-ultra-pure metal.

The metal produced this way has higher energy density than the helium. And, unlike the helium, we can already turn it into energy on earth by… burning it.

Also, harder than burning it but easier than making a fusion reactor, similar levels of energy by shaping the metal into rods, aiming and firing them at a coil of wire on the earth, and using magnetic induction to extract work from the energy gained by falling through the gravitational potential difference.

But it's worse: We don't (yet, or publicly) have useful fusion reactors of any kind, so we can't use this stuff anyway. Tech demos, sure, but nothing useful.

But it's even worse than that: D+3He is harder than D+T, and while it's easier than D+D, there's no way to make a D+3He reactor that doesn't do at least some D+D fusion, and the D+3He reaction rate is never greater than 3.56 times the D+D reaction rate.

But it's worse than that, the only reason to try to avoid D+D in the first place is that half the output is T+P, the other half is… 3He+N. So you get the 3He anyway as a side-effect of a reaction that doesn't require starting with any 3He.

But it's somehow even more ridiculous, because T decays into… 3He.

And and and… if we did have a reactor that could use this, we wouldn't need the moon anyway, because one thing you can do with a fusion reactor is relatively easily travel to the gas giants and extract the stuff from their atmospheres in much higher densities.

(Wish I could remember the original article that had all this stuff in so I could link to it; saw it years back, never found it again).