The South Pole of the moon will end up one giant mega city due to it's constant sunlight. It will be a lot easier to get there once there is even one landing strip.
It "only" takes about 8 million metric tons of aluminium to put a conductive aluminium loop of 1Ω resistance around the moon's equator*, solving all the energy problems that the poles are usually supposed to solve.
While this is way more than we can do any time soon**, that's the kind of mass you'd need to get up there to build the infrastructure for a colony of just high tens of thousands to low hundreds of thousands of people, which is well short of "mega city".
* https://www.wolframalpha.com/input?i=%28%28circumference+of+...
** Last I looked, SpaceX's website was listing 100 million USD per ton to the moon; your guess is as good as mine if or when they'll deliver anything close to the price Musk suggests for serious colonisation efforts
The moon is dangerous because there's no people and civilization is 5 days away at best but if there was already civilization at the moon you wouldn't think it was dangerous.
On top of that the materials on the moon are already "on the high ground" meaning you don't need to spend a lot of money on propellant to get it into orbit. So building space habitats and delivering them into an appropriate orbit on the moon is a tiny fraction of the fuel needed from Earth. To put this into perspective the Apollo Lunar Module only needed 2.2 Tons of propellant to get the upper part of it back into orbit to meet up with the service module. 2.2 tons of propellant is basically nothing with the scales we are talking about.
On top of that if we could produce the propellant on the moon the costs and logistics and difficulty of all of this drop significantly.
So in short the best possible way to lower the risk, cost, and provide functionality is to establish civilization on the moon and get to the industrial age there as quickly as possible.
We're doing it regardless of what you naysayers will say about it because it's the right thing to do for a thousand different reasons. And we're doing the robot thing too. At the same time.
Here's a long list of photos of Earth, a few were taken by humans, most were taken by robotic probes: https://en.wikipedia.org/wiki/Timeline_of_first_images_of_Ea...
One of those strides has been in characterising just how magnificent the human eye, mind and hand are at picking weird shit out of a background.
That said, I see a lot of synergy opportunities for humanoid robots on the surface controlled remotely by humans wearing VR gear.
Why are landing strips the big unlock? Blast effects? Tiny landing legs?
Also the landing strip can be designed to slowly go up hill which could help with the breaking phase as well.
Makes me wonder if you could accelerate to orbital velocity using something like a maglev train and not have to worry about rockets at all.
Rough estimates? Mass drivers make sense. I haven’t seen the numbers for just compressing and leveling regolith.