Also I'd be interested to read how they got around scintallation of the Interstellar Medium.
There is space-based VLBI at lower radio frequencies, with the "Radio Astron" project[0]. That effort works at frequencies which are roughly a factor of ~10 lower than that of the EHT. I'm not aware of any plans for millimeter Space VLBI, but the higher frequency would require higher data rates.
Resolution-wise, the angular resolution is inversely proportional to the farthest baseline, as given by the Rayleigh criterion. [2] To get twice the resolution, you "just" need to double the size of your baseline (and do that in both dimensions, otherwise the angular resolution will be different in x & y). We've maxed out the Earth's baseline, so it seems like orbital radio telescopes are the only way to better resolution. Pretty exciting!
[0] - https://www.jpl.nasa.gov/missions/space-very-long-baseline-i...
[1] - https://asd.gsfc.nasa.gov/blueshift/index.php/2016/07/25/thi...
Going to higher frequency gives you higher resolution for a given physical baseline length.
That may not work for imaging the immediate surroundings of black holes – one can only combine the data for that SAR-style imaging if the source you're observing doesn't vary significantly (e.g., in brightness or flux distribution) between observations.
Why do you say that? The latest High Throughput commercial satellites do 500Gpbs downlink throughput, so shouldn't the transfer time for a petabyte of data be reasonable?
Of course those satellites are configured to transmit data all over the earth, so the technology would have to be used differently for this application.
One of the reasons they've imaged M87* rather than a black hole in our own galaxy was to avoid dealing with ISM scattering - we don't have to look through our edge-on disk, so it's easier to image another galaxy's center somehow. But the Sgr A* image might be in the works already, it was mentioned in both the press-conference and one of the paper's future work section.
The optical equivalent can be dealt with by active optics, or picking frames with best seeing, but for ISM I imagine it is a fair bit different.