I used to work in the amyloid field, which is tangentially related - last I recall the companies were getting to 10~25% the tensile strength of spider silk. Which suggested they were doing something fundamentally wrong.

The most obvious thing, to me, was that they were using a chemical called HFIP "hexafluoroisopropanol" to initially dissolve the silk proteins. This solvent forces proteins into a form called alpha helix which in this case abolishes the protein soluble - but the transition to beta sheet (which it needs to be for its strength) is a poor one. The amyloid field was using this a lot, and the results were horrible. The result was extremely irreproducible results, which the field papered over by doing experiments with an n of 1. My graduate thesis was basically doing experiments without HFIP and getting results that actually make sense (with higher Ns).

I reached out to Refactored Materials, which at the time was trying to do the spider silk thing, to suggest they play with the procedures that I had developed for amyloid work, but a trip to visit SF and give a lecture fell through.

Recently, I was in a conversation about a SF biotech that I think rose from the ashes of RM trying to do the spider silk thing, and apparently they aren't using HFIP (or TFE, trifluoroethanol, a similar solvent with similar effects) anymore, but I didn't get too many more details because these companies tend to keep their techniques close to their chest.

Sadly, I still see papers publishing amyloid work with HFIP and TFE going through my google scholar feed (I am now almost 7 years post-PhD).