I miss the lower 100 ISO on the D100 myself since it is not a sports action camera body per se, but if I had a tradeoff for the D2H I would prefer cleaner, less noisy higher ISO's at 800 and above than to include a 100 ISO. Cannot answer your question about if lower ISO's require a better chip or not.
The minimum ISO rating is set by a combination of two factors in the physical sensor chip:
The quantum efficiency (how many photons are necessary to produce an electron in the sensor well
How many electrons a sensor well can hold without saturating (saturation charge).
Improving the quantum efficiency increases the speed without increasing the noise, so improving quantum efficiency is almost always a good thing (the only exception I can think of is if it significantly increases dark noise).
On the other hand, reducing the saturation charge of a sensor well is generally a bad thing. It increases the ISO speed but at a cost of reducing dynamic range and increasing shot noise. The problem here is that the only practical way to increase the saturation charge is to make the sensor pixel bigger, which would mean reducing the pixel count. However, making the sensor pixel bigger also increases the amount of light that hits the pixel and hence the number of electrons. Thus, changing the area of a pixel will improve the noise but will not change the ISO rating.
Nonetheless, it might still be plausible to argue that under some circumstances it's better to take a 1 MP picture with half the shot noise of a 4 MP picture, except that to a large extent, if the electronics are good, you can do this noise-reduction in postprocessing by averaging adjacent pixels. Thus there would seem to be little reason to issue the D2H as a 1MP camera with incredibly low shot noise.
From where I stand, if the dominant cause of the increased ISO figure is the move to higher-efficiency sensors and there is little reason to complain.
Now all this is true of the minimum ISO that the chip is capable of. To increase the sensitivity above this minimum value, you need to add amplifiers between the sensor and the digitizer. These amplifiers add noise (principally Johnson noise from their feedback resistors). This is why you tend to associate higher ISO figures with more noise in digital cameras. Again, if the base ISO of a chip is increased then you need less amplification and there is less noise for all sensitivities.
One final thing that occurs to me: The D2H has moved from CCD to CMOS technology (their super-secret LBCAST JFET technology). CMOS is plagued with significantly higher dark currents than CCD, which would mean that to control dark noise, you really want to move the sensitivity up in order to keep exposures short. That might be an additional factor. However, as you note in your post, Nikon went to minimum-200 ISO with the D100 which uses a CCD sensor.
Not to take pictures of one's children, particularly when they are small, is a sign of parental indifference --- Susan Sontag
Since Nikon's new technology isn't spelled out in engineering detail, I don't think anyone can predict what the eventual production camera can or cannot do. So far, the test photos I've seen suggest that Nikon has a lot of work to do with noise at ISO 200, let alone 1600 or 3200:
What I was told by the Nikon rep that I talked with, is that there is no improvement in image quality going below 200 ISO, so they didn't include it. Of course that totally ignores people that use low ISO settings in order to do slow shutter speeds for effects like waterfalls, etc...
True, but it's a lot easier to block excess light than it is to enhance insufficient light. Remember ND filters? Contrary to some early reports, the new sensor is MUCH better at higher ISO's than current cameras. It's not the sensor's fault if the photographer underexposes the shot.