N95 masks use a special plastic fabric that can only be produced with really specialized machines. That's usually the bottleneck since it takes months to make and install the required equipment. Also, keep in mind that the production process is already highly automated anyways.
That level of automation is what I was wondering about, how much it would take to remove humans entirely. Admittedly I was working under the assumption that the producer of the masks wasn't also manufacturing raw materials as that wasn't addressed in the article and may be a separate issue in and of itself.
Scaling up means a lot investment but what happens when the demand goes down? You're left with debt from it and little demand. The scaling as such is not the problem here.
Scaling as in just let the machines run longer, normal hours to 24/7 or whatever is necessary, there's more room for a nuanced production if less people are involved right? For clarification I didn't mean scaling as in buying more machines, which I agree would likely require prohibitive investment.
Of course keeping a few machines in the wings is probably cheaper than overtime and unemployment.
[edit] I'd also like to know how difficult it would be to just switch the machine over to making cheaper, non-medical grade masks as demand changes?
It doesn't work with specialized machinery like it does with servers aka "cloud". Machines that are reliable 24/7 workhorses cost money, when the demand is down you cannot share them over the internet so others can do something useful with them.
I suppose it's more realistic with cnc or 3d printing but even then it's not that simple mission, this small company wouldn't most likely try or want to be "aws but for x robots" : )
I suppose if you think about 3d printers, plastic molding, cnc, then we are slowly going to this direction.
While the following articles suggest this very question is an important aspect of robotics that manufactures can take advantage of to mitigate costs and make modernizing less painful
With the following article making a strong case for the re-shoring of jobs that advanced manufacturing options have historically resulted in as well as laying out the benefits these technologies can provide which allow small scale production of niche products.