Two main problems:
First, there is nothing to define "truly ergonomic". There is no standard, nor is ergonomics a field prone to developing a cookie cutter to stamp out standard anythings. As such, when there is no standard, there are far fewer opportunities for economies of scale, which prevents the lion's share of profit margins for major producers.
Second, it would take a lot of research to develop a keyboard which was enough of an improvement for a significant enough user base to justify even a small production run, and those keyboards would then not only have to show profit on their production costs, but also the cost of that research. All at the same time the company would be taking on the risk that they'd created
The Homer in keyboard form and have to take a loss on the entire line. There's just no need to take on that risk when you have a successful product line.
Ergonomics are simply a niche market when it comes to human-machine interface devices, and most manufacturers are quite content with the devil they know. The prevalence of QWERTY is strong evidence of that mentality throughout the marketplace.