In a subtractive colour model you typically have a fourth channel (CMYK) because simply mixing all three primary colours together still reflects too much light to get good contrast.

For E-ink the colours don't actually mix with each other, so some basic maths shows that you would have even worse contrast with only three colours in a subtractive model:

Each ink colour absorbs only a single RGB channel, eg. cyan absorbs red. This means that at each location on the display, 2/3 of the light is reflected even when CMY channels are all at their maximum.

I think E-ink is neither an additive nor a subtractive colour model: it can only produce colours that exist somewhere between the extreme primary colours, ie. it's more of an "interpolating" colour model.

The inks probably come in pairs: black/white, red/cyan, green/magenta, blue/yellow. This gives the expected eight primary colours. For each pair, you would be able to pick any mix, eg. 30% black, 70% white.

For the darkest black you would have 100% black, red, green and blue. For lightest white you would have 100% white, cyan, magenta and yellow.

(This would reflect (1/3) * 3/4 + 0/4 = 25%, and (2/3) * 3/4 + 1/4 = 75% of the light respectively assuming perfect inks and equal coverage for the channels, which is pretty decent contrast)