All right. I have some hard data to offer. In 2018, Alan Edlund, Justin Jones, Randolph Lewis, and Jason Quinn published this paper in New Biotechnology:
Economic feasibility and environmental impact of synthetic spider silk production from escherichia coli
This work was sponsored by the DoE and many will recognize Randy Lewis from Utah.
They indicated that a current generation bioreactor could produce a >75 kDa molecular weight recombinant spider silk protein using E. coli for $761 per kilo. An optimized reactor was projected to reach a rate of $23 per kilo. They use affinity chromatography to purify the proteins. In addition to costs, they also note the quantity of CO2 produced is quite high.
Note that the $23 per kilo is speculating about potential future technologies from a bioreactor several generations beyond what is present now. However, the figure of $761/kilo is, according to the authors, achievable with existing technology.
If you don’t like these numbers, take it up with the authors, the scientists who reviewed the manuscript, and the journal editors.
They do not report purity with their technique, but it does not seem to be an issue. Their separation technique was quite impressive, for me at least.
Although a silk with a molecular weight of around 75 kDa is not as strong as spider silk and, presumably, MS and DS, it is certainly strong enough for normal textile purposes. KBLB looks to have a significant edge in cost that will erode as more research is done to improve protein yields. Spiber has the advantage of deep pockets and pre-existing relationships with big-name clients.