Can spectrum object be lazy?

I noticed that when I pass in a big flux array, it takes a while to create the Spectrum1D object. I am not even doing anything with it yet, just initializing it. Is it possible for it to be as lazy as possible? If so, is such performance tips documented somewhere?

Example usage

Data file is from http://archive.eso.org/wdb/wdb/adp/phase3_main/query?collection_name=MUSE&version=&object===HE%202302-0857&order_main=wavelmin%2Cfilter%2Cexp_start%2Cexptime%2Corigfile&top=10000

from astropy import units as u
from astropy.io import fits
from astropy.wcs import WCS
from specutils import Spectrum1D

filename = "ADP.2016-06-17T18_47_32.888.fits"  # 3 GB, 2 extensions each (321, 315, 3682) float32
pf = fits.open(filename)
wcs = WCS(pf[1].header)
flux = pf[1].data << u.Unit(pf[1].header["BUNIT"])  # This is fast
sp = Spectrum1D(flux=flux, wcs=wcs)  # Takes a long time

xref https://github.com/spacetelescope/jdaviz/issues/2502

Start with the Spectrum1D(flux=flux, wcs=wcs) construction shown in the issue and reproduce it using the linked 3 GB FITS example. Investigate which initialization steps account for the delay and review the jdaviz issue for related context; done means establishing whether lazy construction is feasible and documenting any supported performance guidance.