Why is the B→η′X decay width so large?

A new mechanism for the observed inclusive B --> eta'X decay is suggested. We argue that the dominant contribution to this amplitude is due to the Cabibbo favored b --> (c) over bar cs process followed by the transition (c) over bar c --> eta'. A large magnitude of the "intrinsic charm" component of eta' is of critical importance in our approach. Our results are consistent with an unexpectedly large B(B --> eta' + X) similar to 10(-3) recently announced by CLEO. We stress the uniqueness of this channel for 0(-+) gluonia search.