[Chris is me]

I believe the reason Travis is specifically mentioning next year is that it will be the first game with the new control system where anyone would have any reason to build a 6 CIM drivetrain. Some have suggested that the system is more vulnerable to brownouts than in the past. (I'm not particularly informed on this, I don't know if people have tested this etc)

However, at least in the context of the previous control system, gearing your low gear so that your motors draw <30 amps each when traction limited is the right move. It's a conservative number that should help prevent pushing matches from tripping your main breaker. Be careful to manage current draw throughout the rest of the robot as well.

As for the design, it's not too bad. I would remove the lightening pattern from the front and back rails. The square lightening versus a triangle or X pattern is substantially weaker, and most any lightening pattern in 1/8" wall tube is going to result in a weaker and less rigid tube than a 1/16" wall unlightened tube. And it won't even be lighter! If you want to save weight, you can usually get away with 1/16" wall tube for the crossmembers.

While I wouldn't say this chassis would have "little to no" torsional stiffness, a lot could be gained with a structural belly pan for relatively little cost of weight (since you already need something to mount electronics to regardless). The waterjet or lasered aluminum belly pan is a slick way to do it, but a composite like garolite left solid can also do the trick for a lot less machining time. You could also add some standoffs connecting your two gearboxes together at the mounting holes. Also do keep in mind that your manipulator design may also provide some cross bracing, and in that case your drivetrain doesn't necessarily need to be structurally perfect independent of other subsystems.