most likely you will reduce the load.
if you burst a report it will be queried only once with a large dataset and then sliced into 200 reports, so you will reduce the load on the database (one query instead of 200 filtered ones). the cognos server will also have less load as it will only be processing one job.
however, if memory is an issue here (and how you have configured the virtual memory available to each cognosBIBus process), you may run into trouble when handling the large dataset that is to be sliced into 200 small ones.