The problem is just the size constraints.
Don't be too preoccupied with downsizing right away. The whole point of devboards are to
develop the prototype product first. Their large size is not a constraint here, it make things easier to fiddle around with. Jumpers instead of solders, USB connection instead of serial programmer, power supplies instead of batteries, etc. More importantly it gives you starting point where you at least know things are working, instead of trying to troubleshoot which one of your design decision is not working. It also allows you to figure out what idea works (or not), how things work, what performance/feature/constraint are there, and ultimately figuring out if the solution is right or not for the job.
Only after you able to make your idea works with the devboards then you can move to the next step of downsizing it. Don't jump right into designing the end product especially since you appears to be 1) not familiar with hardware development; 2) not familiar with firmware development; and 3) not familiar with how exactly you want your device to work yet.
The problem is just the size constraints.
I'm not making the bluetooth reach 40km. I'm using the bluetooth to connect to the phones of the users and then using the phones to reach 40km over a server.
That makes a lot of sense then. A phone companion? Or is it NB-IOT? (is it even still a thing?)
Add:
By the way, ISP1507, neat MCM. nRF52 with most of requisite close coupled components in one package. But do you really want to use QFN package with hidden pad? Not exactly beginner friendly, unless you plan to assemble it in a production house.