I hit on the idea when my wife bought a vinyl cutter, which is basically a 2D plotter with a knife instead of a pen. I don't know if anyone makes an affordable cutter that can cut through gaskets (the material is thick and full of knife-dulling materials like fiberglass, metal, and carbon) but if they do, this is a totally doable business. I envision a web site where users can select an existing gasket from their database and have it cut and shipped. Or, if the gasket they want isn't yet in the system, they can draw it out using a vector tool (perhaps tracing over an uploaded photo or scanned image of the gasket). Select the material, provide credit card info and shipping address, and you get a gasket in a week or less.
I wouldn't mind working with someone on this. My strong suit is not coding but I'm okay at project management, research, and architectural stuff. If anyone is interested in chatting about it, feel free to contact me.
The tools generally used are "flash cutters"[1] (the industrial version of your wife's vinyl cutter); waterjet cutters; or laser cutters. They're all CNC controlled. You need a DXF or other vector file to send to these machines. They're often quite user-friendly now, with either projectors or cameras to help position the cuts on the workpiece, and vacuum-beds which make holding down material quick and easy.
SendCutSend (and probably others) do laser cutting on gasket grade cork and ships it relatively quickly. You could probably use them as your fulfillment system doing the work to make the gasket database.
They also make all sorts of stuff out of various metals, paint the metals, thread them, etc. They have greatly expanded my capabilities as a hobbyist who is not a machinist.
One the fulfilment side, gaskets can be laser cut, and there are services which offer this on demand. Batching up multiple orders into a sheet would make it more cost efficient, along with always printing Nx the required amount and then stock the rest.
A website where you can print aftermarket gaskets on demand, and even do it with CAD files or derive them from edge detection in photos would make its money by shipping faster than OEMs, and without worrying about inventory. Retooling an existing custom sticker printing business would do it. Unit margins are like 100x the cost of the paper, and shipping cost margin is the other one.