Agile can work for hardware as well as software. I was part of this in my previous company.
It requires some strict rules and project boundries though.
It probably will not work well for big launch mass production many n clients as first delivery.
What we did is think of a new product, unknown in the industry. Unknown if clients like it or even want it or pay the extra $ for it.
So we choose one client who was going to build a new HQ offered him to participate and actually be a Beta tester and gave him a huge discount. In the next two years we build four versions of the hardware product each time with improvements after the clients feedback and replaced the products in the end with the final product. Production teams were also thinking about cost down improvements in between the cycles.
The end outcome was actually that there were many restrictions and the product was too expensive, so that we did not continue selling it in masses to other clients.
But still it was a succes:
1) because we had four products in two/three years where we normally made one finished product in that time and then hoped the clients would want it.
2) we saved the company a lot of money not going into mass production and failing to sell.
It required more than just agile we needed new faster and cheap production techniques, instead of molds taking a year and €300k a piece, we 3D printed a lot of parts for prototyping and first products.
But Agile requires the entire company to cooperate and understand this WOW. Esp management and product owners had a tendency to interfere last minute with new requirements That is not how it works and really stops the ART. The team has a phase for input, then plans and then leave them the hell alone while they build what you asked for, don't change the requirements in between the synchronised timeslots called sprints and PIs or you get chaos. Yes ofcourse it was chaos the first year
they even had to fire a manager because he could not or did not want to change his personal WOW.