You could parse the files directly using the robot framework libraries.
A small command line tool that takes similar arguments as pybot and then returns a list of test cases broken out into the desired number of groups would be awesome.. While your there throw in OR ability for --include/exclude ..
Here is a small example I started to play with:
#!/usr/bin/env python
from robot.parsing import TestCaseFile
suite = TestCaseFile(source="./Google.txt").populate()
print "suite: "+
suite.name
for test in suite.testcase_table:
print "\ttest case: "+
test.name print "\t\tdoc: "+test.doc
Right now for I just use a Jenkins matrix job to run various browser/os/tags combinations at the same time.