A method for recovering data in a redundant array of independent disks (RAID) is disclosed. In one embodiment, such a method includes implementing a RAID in a storage architecture comprising a RAID controller and a storage enclosure containing a set of storage drives belonging to the RAID. The RAID controller is configured to detect a failure condition associated with at least one of the storage drives in the set. In response to detecting the failure condition, the RAID controller sends, to the storage enclosure, an instruction to copy data from a first storage drive in the set to a second storage drive in the set as part of a RAID rebuild process. The storage enclosure receives the instruction and executes it without substantially involving the RAID controller in the copy process. A corresponding system and computer program product are also disclosed.