I did my PhD on exactly this area (thesis on yousefamar.com, DM me if you'd like to chat about gamedev and netcode!) and optimising for cost and easy setup to support indie devs.

Afterwards, when I tried to validate this as a product (libfabric.com) I realised that I'm trying to solve a problem that nobody has, except for a very small niche. The main problem that developers of large-scale networked games have is acquiring players. They use whatever existing service+SDK (like Proton) for networking and don't think about it. Once they have good traction, then they can afford scalable infrastructure that others have already built, of which there are many.