This going to be one of two things IMO,
#1 slippage of couplings/pulleys or friction due to mechanical wear/failure.
#2 Lose or breaking/ trapped wire

However, everything Neale mentions is correct and could very likely be a cause of trouble and most certainly will be straggling performance, so I 100% agree with him that you should up the volts for the Nema 34 and go with a toroidal type supply. But first I would try to identify the cause of this issue if it as been working ok for a lengthy time, the simplest way to check the PSU would being just running the X-axis with the other axis unpowered.

My suggestion would be to write some G-code with 1000's of short moves at full rapid speed but just long enough to let the motor reach full speed before changing direction. Zero and Mark the axis before starting the run and finish by having it come back to Zero.
If it returns to zero, then repeat but this time run the other axis at the same time, if it then drops steps then it's a power issue.!