It’s the computer telling the drive “write everything to durable storage (as opposed to some kind of in-drive cache/RAM) and tell me when it’s done”.

After that command it should be 100% safe to pull the power because everything SHOULD have been written to flash. That’s the point of the command.

It’s interesting that the drives that do it wrong still take time indicating they’re doing something.

The DRAM cache does not hold user data. It holds the flash transition layer that links LBAs to NAND pages. Higher performance drives use 1GB of DRAM per 1TB of NAND. In cheap DRAM-less drives if the I/O to be serviced is not cached in the 1MB or so of SRAM it has to do a double lookup. Once to retrieve the full FTL table from NAND and a second lookup to actually service the I/O.