Rendering limits in Fusion?

Evnin'

As I am doing user interface use cases I started with Resolve for manipulating shots and added effect inside Fusion to display user interface elements like LED glow or OLED displays...which turned out nicely...

Now I started to do visual designs inside blender for user interfaces which involves DLP beamers via mirrors...which turned out only one render software could do the job...

How is the 3D renderer inside Fusion compared to 3D render software out there?
Are there any limits/drawbacks when it comes to special lighting conditions?

I assume Fusion isn't capable of doing bidirectional raytracing like luxcore does?

Fusion's 3d renderer is very basic. It has no raytracing capability at all. So no GI, no inter-object reflectivity (you can make a spherical reflection environment but it's difficult to make 3d objects reflect one another), no IBL, no SSS.

The advantage of the stripped-down feature set is that Fusion's render times are measured in seconds rather than minutes or hours. But Fusion definitely is not a replacement for a proper 3d program.

That's why I saw so many people using blender for 3d compositing as it also features the opencv tracker like Fusion does...

Can the camera and tracking data be exchanged between blender and Fusion?
Blender has a nice feature of rendering layers so you can just overlay the shadows onto a video or picture...

You can export a Blender camera using Alembic and import it to Fusion with File > Import > Alembic scene. Transferring a Fusion camera the other direction is more difficult. I think you can send it out through FBX, but I've never tested the workflow in that direction.

You can send animated geometry to Fusion with Alembic, too, but you'll have to re-shade it, and maybe retexture (but the UVs will be fine). Not recommended if you use UDIM.