Without Domino the pipeline reads the files directly, and it runs in two stages that reuse machinery this forum has covered separately: NSF to PST, then PST into 365 by whichever door fits the volume.

Stage one, the conversion: an NSF converter reading the database files standalone, SysTools NSF Converter as the standing example here, walks each archive and writes PST preserving the Notes folder structure, rich text bodies and attachments. The Notes particulars from the reading thread apply identically: individually encrypted messages need the original user ID files or they skip with a report, and any names.nsf in the pile is an address book converting separately as contacts rather than mail. Convert into one PST per original mailbox, the mapping that keeps stage two sane.

Stage two, the import, sized by volume: a handful of archives go the manual road, each PST attached in desktop Outlook beside the target 365 mailbox and dragged in, verified live. A pile goes through the compliance center's PST import service, the network upload with AzCopy and a mapping CSV assigning each PST to its target, with shared mailboxes or inactive mailboxes as sensible destinations for departed users' archives, searchable without burning licenses, the same reasoning as the OST import thread linked alongside.

The verification pass closes it: item counts per folder against the converter's report, a spot check that attachments open and Notes rich text reads acceptably in Outlook and the skipped encrypted items list handed to whoever owns the legal question of whether hunting the old user ID files is worth the trouble. Direct NSF to 365 tools exist collapsing both stages into one, worth the look when the pile is large enough that the PST middle step becomes its own storage project.