I noticed I slightly misquoted the Epson pre-release brochure quote. It should say the UC D6 prints have "lightfastness ratings over 80 years."
The early discussions of Claria being a hybrid of dye and pigment are, I believe, wrong. They are dyes. What makes them different is that they are more complex molecules that are stronger. I believe the chemistry is coming from Fujifilm. The Fujifilm descriptions of its dyes varies between "azo" and "novel azo" dyes.
The Epson claims remind me a bit of their "Archival matte" paper name -- a rather inappropriate name for an acidic, quick-to-yellow paper.
While I think the UC D6 K MSDS was simply a mistake, I do note that Epson is not without some support for claiming the 80+ year life for their dye images, on appropriate paper and when properly displayed and stored. They probably base their claims on Wilhelm's test methods. These allow more fade and color shift than most of us would be happy with for our B&W prints, where such things show up more. Additionally, the Wilhelm test method apparently misses what is happening in the deep shadows, where the Claria has a weakness. This shows up in the http://www.aardenburg-imaging.com/ testing of the color Claria inks in the E5 patch. However, using the Wilhelm approach, the Claria inks can test as equal to the UC pigs that were used in the Epson 2200.
It looks like it will take fade testing to see if there is any difference with the D6 inkset, and with the evidence I've seen so far, my plans will assume there has been no change. I will continue to use the Noritsu dyes, but shift to using considerable color to build the grays, balancing the different directions the K and color-generated grays color shift. I may also do an informal fade test (unless Epson goes the unlikely route of responding with more information).
Paul