World Building and AI in Space

One of the best bits of writing is what is called "world-building", when you create pictures in your head of how the future might look, how it might work, both from an engineering but (as is often overlooked) financial basis. Space is hard and expensive, and both angles should really be considered.

You can't just invent a space colony - the key question is why? What are your characters doing there that makes all the effort and risk worthwhile? And then additional questions come along, such as what are the implications?

For a scenario with million AI data centres I might start wondering:
- With that quantity, would it not be better to manufacture off-world, maybe on the Moon (as some have already proposed)?
- Given the energy source (the Sun) can be eclipsed by the Earth, wouldn't further away be better?
- Given latency to the AI is critical, shouldn't the human population using the AIs be close, which means off-world?

So the result is wider than the initial proposal: a space based economy and extensive human presence.

We live in interesting times, with all that comes with that saying.