I think the main problem FSR is going to have is convincing developers to use FSR instead of their own solutions. The reason why DLSS 1.0 didn't get much traction when it released was because many game engines already had their own up scaling solutions that were better than it. So while FSR will work on any GPU, that doesn't really matter because engines that have their own upscaling technologies (which most do) already work on any GPUs. The reason why DLSS is getting mass adoption right now is because DLSS 2.0 is massively better than every other upscaling solution in the market.
Remember, it's not about the performance that FSR brings because every Upscaling solution on the market will give you similar performance... It's going to be about the image quality vs current solutions. That's why DLSS 1.0 failed and if FSR can't bring better image quality than current upscaling solutions like Unreal's Temporal Super Resolution, I doubt it will gain much traction.
So we will see how it goes but so far, it's not looking good going by Anandtech's article.
PC Specs: CPU: 7800X3D || GPU: Strix 4090 || RAM: 32GB DDR5 6000 || Main SSD: WD 2TB SN850
