Considering a small 6S battery could be onboard the module along with the controller thus eliminating basically all the mechanical drive parts they don’t already stock. The idea in general is very feasible, for a robot printing the module frame could lower end user BOM from trampa to very low cost.
Then you need a central brain that also powers the modules rotation, sending drive signals. Bonus points if it has an automatic swerve module battery changing mechanism.