I have OpenGL code running on FreeBSD, Linux, MacOS X and Windows. I'm not going to write a second implementation for Apple systems. Vendor-specific graphics APIs are a throwback to the '90s. I am categorically not going to support Apple's Metal API; they need to bear in mind that much software using OpenGL is not Mac-specific, and given the poor state of Mac graphics Macs are decreasingly a second class platform for serious graphics use. Even FreeBSD has better OpenGL support(!)

"Market power" is fairly meaningless when you support multiple platforms. Most games ported to Apple platforms use OpenGL; using Metal would restrict their market significantly. And expensive high-end software has better support on platforms where you can use expensive high-end GPUs and compute, and that's not today's Apple hardware. Metal is a non-starter for a very large number of people, and they need to get their OpenGL and Vulkan support up to par or many of us will cease supporting Apple platforms entirely.

This is shortsighted attitude that, in the long run, will be toxic for Apple. It's one thing to not give developers what they want because you can't ( doesn't fit the technology or the product, don't have the resources or whatever). It is quite another to do that simply because you won't (e.g. for strategic lock-in reasons).

Yes, the iPhone gives them all the market power they need with some developers, even many devlopers. But others don't care about the iPhone for various reasons. Those developers, grumble, moan and eventually buy computers that aren't Macs. And when those developers make great things they're outside of Apple's ecosystem, encouraging other developers and customers to leave it as well...