Correct, still about 500ms to get decent levels of separable reproducible grayscale. Claims of 10Hz like Dasung's A2 are often marketing interpretations of unstable 1-bit states. This is a limit of the physics of electrophoresis. To make it faster, they would have to abandon bistability. Some startups have gone that route, eg: ClearInk but they have not yet reached commercial scale.
Clearink does not rely on the foreground pigment particles coming to "rest" at a fixed position relative to the background pigment, ie: achieving a stable state. Instead they rely on driving the field for as long as the foreground is needed. As soon as it decays, the background pigment instantly rises back to the internally reflecting top plane. This allows the display to achieve much faster speeds since the pigments can be optimized for speed rather than stability.
From what this description sounds like, the foreground pigment would only remain at the forefront so long as a charge was applied. Doesn't this defeat the energy savings of eInk in that once you have updated the display it doesn't cost anything to hold its image?
Yes, it would lose the image once the display is no longer driven. The main attraction is the removal of the need for a backlight and the sunlight readability and all the other advantages that come from using actual pigments.
