Santa Fe's last 4-axle high horsepower GE units, the B40-8W, weigh (acc. to Greg McDonell's "Field Guide to Modern Diesel Locomotives") 289,000 pounds: close enough to 72k/axle that maybe nobody quibbles. Take that as a starting point. I think the 12 cylinder GEVO engine weighs almost as much as the 16 cylinder FDL. (The original design of the HDL-16 was only 10% more than the FDL-16, but part of the cure for the HDL's problems was more cast iron in the engine frame: lots more. I don't have documents to back this up, but I have a feeling the GEVO has inherited the extra iron.) That's the engine: to be EPA compliant, ES-series locomotives have extra radiator/heat-exchanger gear: this probably adds up after a while.
---> Intermediate conclusion: trying to get a modernized version of the B40 might go over the permitted weight even before you think about replacing DC motors with AC.
AC traction motors are lighter than <i> equivalent capacity </i> DC motors, but the GEB13 is a much higher capacity motor than the DC 752: I'm sure someone here can give an exact figure (please, someone?), but at a guess it may be a bit heavier than the 752. And the above-the-frame part of an AC diesel electric is heavier than that of a DC motored one: after all, you have essentially the same alternator and rectifier set up, and then have to add to that the invertors to transform the DC into variable-frequency AC.
---> So, a straightforward "B40-ES-AC-W" would almost certainly be too heavy.
Could a B-trucked design be gotten within the weight limit? Well, you could omit the fuel tank, and couple a -- maybe we could call it a "tender," if that doesn't sound too steam-era -- to (or between a pair of) the locomotive(s). But that just moves the unpowered axles we wanted to get rid of to a separate vehicle! (And ATSF's partner in the merger that formed BNSF did try out fuel tenders, and seems to have decided that they were more nuisance than they were worth.) The other possibility would be an all-new carbody design-- think "Freight Genesis" -- with a lighter load-bearing structure than a road-switcher frame. That, however, would involve BIG MONEY: maybe it will come, but not on (what I called) a "proof of concept" prototype.
On another point: I don't think 25 units is too many for an experiment. GE fielded 50 preproduction GEVO units in railroad colors (5 to UP, 30 to BNSF, 15 with DC motors to NS) for testing before the production introduction of the ES series at the beginning of 1985. You want a test to provide meaningful statistics, and to be extensive enough that it has a fair chance of picking up a problem that doesn't occur every day. And BNSF's total locomotive fleet is up towards 7,000 units: as a proportion of total fleet size, 25 locomotives for BNSF is less than one prototype warship for the U.S. Navy!