Cobol (yes, really!)
Tab-sepearated and Comma separated text (CSV) files
Various types of config files, OS scripting and logs
(related) Very Large text files
In some of these cases, the lack of flexible tabstops is a real PITA, especially with Tab-separated columnar data as produced in a number of logfile types and other scientific data dumps. Brief (one of my old favourites), allowed setting tabs with a command like "tab 5 20 50 55" which would set tabstops at each position, with the remaining tabstops defined by the last two (ie. 60, 65, 70 etc.). That and virtual space are probably the biggest asks for me in Sublime, and are probably the reason most of my colleagues stick with their favourite editor instead of switching. That and the restrictions in macro flexibility. Several TM users have mentioned to me about Sublime's code folding ('dumb' indent based and not observing tmLanguage apparently) and indenting style limitations; neither bother me personally but it comes up regularly in conversation.
On the last item above, large files; Sublime seems to struggle a bit with this, especially with complex manipulations and undo. The other main editor I use is Crisp which performs superbly in this regard, especially when applying large S&R ops with regexes and column manipulations with rectangular selections. Files could be over a gigabyte in size. I have not yet tried ST3 with such files, could I expect any improvement there?
I'm spreading the word, but Sublime's a hard sell unless in addition to its niceties it fulfils bread and butter requirements for a given user. But I do like what's gone into ST/2, and recent developments. Keep it up!