Good afternoon!
If I had an unlimited budget, I'd probably go far overboard.
Hire a team of engineers, designers, researchers and other useful and talented people.
Then give them the base instructions/idea and let them have at the arising challenges. Again note the unlimited budget, thus the size of the team wouldn't be limited by financials, nor would time be a huge bottleneck, as cost-by-time would be negligible when dividing by infinity.
Indeed communication would be a bottleneck as the number of people goes up, but I notice myself digressing.
I'd say if I had unlimited money to spend on the project I'd find people passionate about creating a new design, and basically reinvent the wheel. Finding every problem on the way again, and try to find other ways to conquer/solve this.
I'd aim for large-scale, high-precision/high-accuracy, high-repeatability, low-maintenance, multi-material and multi-head. (And if possible at a realistic product-price and at a decent-speed.)
I'd suggest something akin to a planet-builder, but I see that's indeed far-fetched.
Semi-realistically I'd aim for a print-volume of 1~9 m3 and/or up, with multiple heads and/or materials with 0.1~0.01 mm x/y-precision, and if possible not layer-based.
And realistically (as a student) I'd say I'd build one myself rather than an existing one, I know this means a higher cost and finding and solving problems yourself.
But I (personally) find this part of the fun.
(As well as the fact that you can mix and match specs to your heart's/project's desire)
probably FFF, ~25x25x25cm build-volume, double- material or head, remote-driven-direct-extrusion, CoreXY, 32-bit, and other things as I run into them...
TL;DR: Build one myself
Best greetings and a pleasant day to you!
~Sjaak