Initially, this was quite a problem with version 1 of the AirSpy.  I opened it up, scraped paint on the inside at both the USB and SMA connectors and copper taped/soldered (where I could) bridges from each connector to the now bare aluminum (no solder there) and from each connector to the PCB ground plane.  I also installed ferrite with two to three turns through each bead on the USB and RF input cables located at the AirSpy, itself.  I found that even the non-ferrited antenna input cable picked up USB noise and coupled it into the AirSpy.  A lot of that has changed with Version 2 which I now have and especially the HF+, but still have trouble with my ISP (Rise Broadband/ Jab Broadband) getting into the receiver.  I have very aggressively (and I do mean aggressively!!) loaded the RiseBroadband cable with large heavy ferrites with multiple turns through each.  When I'm serious about radio astronomy, especially with the ham - it - up, I turn power OFF on the ISP hardware.  My AirSpys have locked as low as 17 MHz.  I believe because I'm pushing that, I get more vertical lines, but I've learned to live with that.  I can comment that the AirSpys are head-over-heals better than the 'competition', SDRPlay which I also have.  It is absolutely miserable for vertical lines, overload, and just 'garbage responses' in the lower frequency ranges and does NOT respond in a dB manner.  I will NOT be spending any more $$$ on the SDRPlay products ! ! !   Besides, I vehemently dislike SDRuno ! ! ! !

I presently have the AirSpy set up on a crude interferometer (two antennas) on 432.050 MHz (for Cygnus A).  Here (and attached) is a screen capture of my present condition with the AirSpy terminated at the AirSpy with a 6 dB pad (12 dB return loss or better than a 2:1 SWR).

Regarding the USB noise, just be thankful Firewire died.  It was TERRIBLE for radiating RF noise.