This is something I have discussed with Steripen over the years. What is needed is a UV-C LED radiating at 254 nm. So far, we (the science community) can make LEDs in the UV-B range at about 350 nm (from memory). Progress to a shorter wavelength is slow as it is hard to make the required semiconductor material. 254 nm corresponds to a huge band-gap, and we don't actually have any obvious candidates.
Would a 254 nm LED be better than the Philips germicidal discharge tube used at present? (That also radiates at 254 nm.) It would be hugely better! I was going to say 'infinitely better', but I couldn't quite manage that. :-) The power consumption would drop massively.
We don't try to use UV-B photons for two reasons. The simple one is that the UV-B photons do not as much energy as is found in the UV-C photons, and you need a lot of energy to damage the DNA connections. The more complex reason, and the major one, is that we actually need photons at ~254 nm in order to break certain bonds in the DNA. If the photon has an energy significantly different from that, the bonds don't respond. It's a resonance thing.
It will happen, just not this year.