I think it is unlikely that the board is experiencing a failure since, as mentioned above, the board is capable of scanning at rates greater than 37003 with other LabView VI's, just not with this one.
And, ironically, the problematic VI is actually based around the Continuous Acquisition and Chart (buffered) VI, and it seems to be that VI (specifically the AI C-Scan VI in it) which is causing the problem.
Can you think of a reason the hardware would fail under the AI C-Scan VI but not under the AI Mult. Pt. VI?