No, I think it is wrong.
Your calculation is just theoretical output calculation for 2V and 4V to 32ohms, but actual current hits limit in low ohms and voltage gets down.
V=IR
I = V/R = 2.0/32 = 0.0625A (or 62.5mA)
W =IV = I*I*R = 0.0625 * 0.0625 * 32 = 0.125W = 125mW (your calculation)
This is theory, “IF” SR15 can output 2.0V in 32ohms without hitting current limit.
You can see following guy measuring actual output voltage before distortion.
https://sandalaudio.blogspot.com/2018/08/astell-sr15-dap.html?m=1
The site is not mine but he is famous (and probably only one real meaningful reviewer) in Japan.
and the graph for unbalanced output (peak to peak volt):
https://1.bp.blogspot.com/-yaWFcV5Q...fYYh8Ii860zIh0d3hZ3bmoFwCLcBGAs/s1600/012.jpg
It is peak to peak voltage, so divide it by two and divide again by 1.41, or just divide by 2.83 to get voltage RMS.
So in higher impedance, it is near the theoretical output, 5.65V peak to peak, which is 5.65/2.83 = 2.0V rms.
However, you can see it significantly drop in low ohms.
It looks like 3V peak to peak for 32 ohms in unbalanced, thus:
3/2.83= 1.06V rms
V=IR
I = V/R = 1.06/32 = 0.033A (or 33mA)
W =IV = I*I*R = 0.033 * 0.033 * 32 = 0.035W = 35mW
This is just rough estimation read from the graph measured by one reviewer, but it matches my feeling.
SR15 (or SA700 which both I have) in low impedance like 32ohm or 16ohm is weaker than other DAP like AP80Pro which I also have and compared.
A&K DAPs are really nice for efficient IEMs or high impedance and efficient headphones, but not for power hungry low impedance headphones (like MrSpeakers Ether).