Is this true for text? Glyph rendering is complicated and I have a hard time imagining a high quality GPU renderer that can beat a CPU renderer due to this.
I believe the situation has changed in the last year or so? I remember reading a bunch of stuff about how text rendering is finally being able to be pushed onto the GPU lately.
The SDL3 TTF library has the TextEngine API to build a font atlas targeting the CPU or GPU (eg https://wiki.libsdl.org/SDL3_ttf/TTF_CreateGPUTextEngine). Its slightly more cumbersome because you have to work with TTF_Text objects instead of char*s but it's quite doable.
