First images from Euclid are in

A telescope that could zoom into an exoplanet would have an f value of a kajillion or so.

Whereas, multi-site telescopes spread across the Earth have already been demonstrated as a feasible technology (recall the black hole images). It is well within our ability to set up a constellation of satellites, perhaps spanning a few of the Earth-Sun Lagrange points.

You "just" need to get far enough away (~600AU). Interferometry is extremely difficult to pull this off with and it’s further complicated by the host star being so much brighter than the exoplanet.

See this recent Fraser Cain interview with Dr. Slava Turyshev:

https://www.youtube.com/live/lqzJewjZUkk?si=WWNdR1PESYzD0d4X

> you'd have to get really far back from the Sun to resolve the image, no?

Yah, a few hundred AU.

> you are beholden to the orbital mechanics of your viewing satellite as it plods along.

Yah, any mission like this -- interferometry or gravitational lensing -- is going to be super long and hit very few targets.

> Whereas, multi-site telescopes spread across the Earth have already been demonstrated as a feasible technology

Yah, at radio frequency while pinned to a common rock. The wavelength of visible light is hundreds of nanometers and we're talking across massive distances and significant gravity gradients and even relativistic corrections. The "big" space interferometers currently being considered are in the mid-infrared (e.g. longer wavelengths) across baselines of hundreds of meters.

All of these ideas are really hard.

So why not use interferometry instead? Well, it has some significant drawbacks. For example: the Event Horizon Telescope used radio telescopes - and pretty much had to, due to how interferometry works: you need to be able to compare the phase shifts between the multiple telescopes, which means you need to be able to sample the signal faster than the radio frequency you're using and record it. The EHT records 64 gigabits per second for each telescope, and then all this data needs to be combined to compute the resolved image. This amount of data would be problematic for space-based telescopes - even on Earth, it was not practical to send multiple petabytes over the internet, so it was saved to hard drives which were shipped by truck instead. This isn't practical in space, so you would need to transmit the data by radio, which means you'd end up with some crazy ratio of thousands of hours of transmitting for every one hour you spend recording.

I just presume that such observation spacecraft makes sense to be made as a rather long term investment than some one-time fly-by thing like New Horizons, so the sensible thing to do is to have it in an orbit around the Sun at that (≥650AU) distance and to take multiple observations. Of course, the easiest solution may be for the orbit to be elliptical one, with the useful position only around its aphelion, in which case it will be able to observe only a single target indeed. But, I also presume that past the first few such elliptical orbit experiments the aim will shift to have the spacecrafts set on distant (and most likely circular) orbits that make possible observations at any time and thus its target would be something akin to a horizon rather than a single point, wouldn't it?

The other nice thing about being on an escape trajectory is that because of how gravitational lensing works, rather than a focal point you get a focal line. So as the probe continues moving away from the sun, it can continue imaging its target (and the image quality may get better as it gets further out, since the ring gains more separation from the surface of the Sun.)