I ride a bike to work every day. I’m never sweaty. The infrastructure to cycle exists so I won’t get run over by cars.

You’re not listening. It’s not that it’s hard (although it definitely is), it’s literally just infeasible financially and time wise. You cannot spend millions developing an engine unless you are a large AAA studio. You can’t pull up your bootstraps your way into making a modern game engine within the budget you have to make a game.

As for Godot:

1. While games like Domekeeper and Luck Be a Landlord are great, they are made by two people and one person respectively. It has not proven itself as an engine capable of supporting the type of development cycle and team necessary for larger projects.
2. The best games released in Godot are visually vastly inferior to anything you can whip up in other commercial engines. Its focus has been on 2D, and the 3D games released in it don’t look great. Users expect more from bigger budget games.
3. Godot is very new. Many games started development in its infancy, and some before it was even released as open source. Not to mention that most studios have existed much longer and are already established in an older engine, with lots of capital and knowledge locked up in those softwares. There is a lot of inertia to adapting new technology.

It’s not “the easy route”. Making a game engine is a tremendous investment these days. If you are making anything other than a game that looks like early 2000s or earlier, you need a pretty capable engine that takes years to develop. That’s on top of the time it costs to make a game, which is also typically years. Not to mention that your proprietary engine will have subpar tooling and make your game development slower.

For anyone but industry giants it’s not feasible to make a modern engine. Unless your game is not aiming to play and feel like a modern game, you have to run with an off-the-shelf engine.