Describe the issue:
I found that dask.nanstd will bring different results with numpy.nanstd as follow.
I think this may be a bug in dask.

Minimal Complete Verifiable Example:

import numpy as np
import dask.array as da

print(np.nanstd(np.array([np.inf,np.nan,-np.inf])))
print(da.nanstd(da.array([da.inf, da.nan, -da.inf])).compute())

Output:

nan
0.0

Anything else we need to know?:

Environment:

• Dask version: 2025.4.1
• Python version: 3.10.0
• Operating System: Linux 6.11.0-25-generic Ubuntu 24.04.1 x86_64 GNU/Linux
• Install method (conda, pip, source): pip
• Numpy version: 2.2.4