Looks like the axle was a weld up job...
Original axle cut to length, then one end of the axle was bored with a hole, the other end was turned down to fit into the other end for alignment. Two ends of the axle matched together and welded at the seam..
As mentioned countless times, welds unless done in a very highly controlled condition with proper post welding heat treatment will have different hardness across the weld, cracks, stress risers, gaps, gas bubbles and numerous other defects waiting to cause a weld failure. This is why welds are the most difficult and complex method of making a joint as there are SO many items that can go so wrong.
The axle failed at the welded seam.. absolutely predictable and why welded axles are a hazard and can be lethal. Image for a moment if that axle were to come apart at 100+ Mph, full throttle (which is the moment they are most likely to fail.
Do this properly, get an axle made with the the proper lengths of the proper material (8640 alloy steel), heat treated to the hardness required (Surface hardness not higher than 50RC with a soft core for impact and load reversal endurance), splines ends with precision fit. The stock snap ring location is sort of ok, but not ideal.
Or
Get another one of these axles that were modified by welding, cut to required length, then sent the axle out to a speciality shop that can cut the proper splines and snap ring groove on the cut off end. We did this on the rotary LeMons racer, never had a problem.
BTW, axles are considered a consumable item on real race cars as they twist, take a set and will fail due to fatigue with race hours. Indicator for this is a painted line on axis with the axle to indicate axle twist and logging of race hours per axle. Once the number of race hours is up, the axle is replaced with a proven (test-certified) good new one.
On a power train with any significant amount of torque, axles are a serious failure point. Keep in mind it is not hp alone that snaps axles, torque and tire grip is what often snaps axles.
Bernice