Frustum culling makes it so that the graphics hardware doesn't need to work at all on things you can't see. There is nothing built in like that, because neither XNA Framework nor the graphics driver can know what the vertices will look like after your possibly custom vertex shader has processed them, so there's no way of knowing what the bounding shape of your data would be when rendered. You, however, have this knowledge, so you can do frustum culling if you want.
However, there are other kinds of culling. The hardware does back-face culling, where triangles facing away from the camera are not drawn. The rasterizer also does guard band clipping, where triangles (or parts of triangles) that are outside the visible area of the screen are not rendered outside the screen. Finally, there is depth clipping, where triangles in front of the near clip plane (hither), or beyond the far clip plane (yon) are not rasterized. If your objects have a high rasterization cost (i e, "fill rate cost") then whether the object is on screen or off will change the frame rate significantly, even if the transform hardware has to transform the object even when it's not visible, because transforming vertices is often quite cheap compared to filling pixels.
