I'm one of the founders, I exited after the acquisition and now work in VC.
It's absolutely possible, the custom EPD was $40USD, we used injection molder keys, and a delrin keyframe. The remaining components for the display timing controller and the keyboard controller were dollars each at scale (excluding the TCM chip which was relatively expensive).
The entire BOM cost of goods was far under $100, and was comparable to a Macbook butterfly switch keyboard module.
The entire design with Foxconn / it's customers was specifically for MOQ's of 300k units. We did beta units to 40+ countries.
The NRE for the display was also double your estimate and took around 9+ months to get finished + a few months of delays due to terrible yields on the first Mass produced units.
I'm one of the founders. The site crashed due to the traffic, we sold the technology, IP and some of the key talent joined Foxconn late last year; and scaled back our servers during the transition.
I'm a keyboard enthusiast, it started with an IMB model M and that's how the original prototypes started. The feedback regarding adopting Apple design and ergonomics was a reflection that our main strategic partner, investor, and eventually owner was Foxconn, and the customers we engaged with were B2B and set the spec.
For our Macbook projects, it was specifically a Butterfly switch spec we have to match in performance and design. I'd have loved to create a product aligned to what you're referring to, but strategically we have to achieve a MOQ of displays to be economically feasible, and our cooperation with Apple was easier to secure the 300k of displays required than trying to tackle the B2C premium keyboard consumer market.
I'm one of the founders, E Ink displays have their display drivers on their glass substrate, meaning you would have a border that on a single key display would be too significant. We went with one large display with only one driving board for the entirety of the keys. This greatly reduced the unit cost to have one large EPD than 80+ small ones.
I'm one of the founders, the first unit used a prototype EPD from E Ink Holdings, a custom display with around $500k in NRE was developed for a Macbook size keyboard module.
This single large display with a magnetic mechanism operating around the footprint meant the total unit cost in COGS was significantly lower than $100USD at scale.
The entire mechanism was only a few mm in height including the display. The travel and tactile profile was designed to a butterfly switch mechanism spec.
I'm one of the Founders at Sonder.
We found from our test units that uses can work up to 30% faster with dynamic keyboards and had strong responses particularly from developers.
The company, the intellectual property and some of the key talent was acquired by Foxconn, the current design is quite polished, as is the user interface.
In regards to the cost point, it's one large display is a magnet mechanism operating within the periphery. This greatly reduced the unit cost and complexity of the design so it was scalable and comparable in price to an existing laptop keyboard module (when supply chain costs and keyboard layout revenue factored in).
The entire design with Foxconn / it's customers was specifically for MOQ's of 300k units. We did beta units to 40+ countries.
The NRE for the display was also double your estimate and took around 9+ months to get finished + a few months of delays due to terrible yields on the first Mass produced units.