SVG is resolution independent. A filled circle or a font glyph inside an SVG is described mathematically rather than as a grid of pixels, so it renders sharp at any size, from a tiny favicon to a full-screen banner. Rasterizing to AVIF bakes that vector description into a fixed pixel grid at whatever scale you choose, and from that point on it behaves like any other pixel image, sharp at its native size and softer if it gets displayed larger later.
Choosing a scale is a one-way decision
The scale setting decides how many actual pixels the AVIF output ends up with. A 1x export of a small icon SVG might only come out 64 or 128 pixels across, perfectly fine for its original spot but blurry if reused somewhere larger later. Exporting at a higher scale than you think you'll need costs little, since you can always resize a large AVIF down, but you can't add resolution back into a small one after the fact.
Interactive and animated SVG features don't survive rasterization
SVGs can carry embedded fonts, CSS animations, hover states, or even JavaScript interactivity, none of which carries any meaning once the file becomes a static image. AVIF output captures a single frozen frame of exactly how the SVG looked at the moment it was rendered, so any motion or interactivity built into the original file is gone entirely in the result.
Gradients hold up especially well in AVIF
SVG logos and icons often lean on smooth gradients or subtle glow effects, and AVIF's higher color bit depth handles that kind of shading better than PNG or JPEG, avoiding the visible banding that lower bit-depth formats can show across a wide, soft gradient. That's one more reason AVIF suits gradient-heavy vector designs particularly well.