The lede is sort of in there, but buried - or at least not talked about from an economic perspective:

Right now, women consumers put up with one-ish body type (although fit model shapes vary by brand) - manufacturers thus make up to ten sizes or so of any given garment. Google fashion industry waste if you want to learn some depressing things, but - because of fashion lead times, production methods, etc, a lot of these clothes will not get sold. So there’s production waste.

There are roughly 10 core body types according to this website. So, to make ‘properly’ shaped garments for a much larger group of women is going to take roughly 120 different garments for a single design.

This just isn’t going to work for manufacturers given current production methods. I’m working on a fun sweater company right now, and it’s a very analog process - with humans and production and yarn all in different countries - ending in a single garment for analysis that is then put on a model for photography. I cannot imagine trying to scale it to 100 different shape/size combos.

Upshot - right now: choose from the following:

1) create mild differentiation and hit a product target that blends looking good on the site/shelf/model with one that looks good on the customer; keep 90% of the market

2) lean in hard on one of the “10”-ish body types - give up the rest of the market, but have happy customers

3) Try to sell stuff that can get auto-sized properly via algorithm and delivered “on-demand”

Most big companies are big, and therefore chose 1. Some smaller companies chose 2. In the happy circumstance that they chose 2 for conventionally attractive bodies, you’ve heard of them (Chanel). Some have transitioned into this space over a longer period, like Burberry. If you’re not a target customer, they may still have fans, but you might not have heard of them, e.g. Good American.

A few companies have tried 3 — direct to consumer via brick and mortar retail — (there was an MIT company deploying Shima Seki knitting on Newbury street in boston years ago), but they inevitably seem to move to a fast deployment D2C shipping model.

I think this is likely because if you go into a boutique you do not want to pay $600 for a garment and then have to wait three hours for it to get made. Online this feels more palatable.

So, we’ll probably see some continued innovation on the robot-knitting side of the world over the next ten years. In the meantime, companies mostly do what makes economic sense. And, it’s worth noting, operating the “automated” knitting machines and designing for them is no joke, it’s hard — really hard, and the software can be abysmal. So, this is an industry that’s a long way from rapid change, at least right now.