Dual mining technologies on a single chip is not a good idea, there are almost no common parts between the two hashing engines, the silicon for the scrypt engine is a waste of space unless LTC price goes up considerably. Also, the scrypt engine will be relatively power hungry.

For those of you unfamiliar with scrypt physical implementation please read this excellent article by Alpha Technology:

https://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0CCEQFjAA&url=https%3A%2F%2Falpha-t.net%2Fwp-content%2Fuploads%2F2013%2F11%2FAlpha-Technology-Scrypt-Analysis-on-FPGA-proof-of-concept.pdf&ei=ljMXVarbNJHaaIH-gZAB&usg=AFQjCNG9nKGtWnMgUeKH2mHZclPgyOxKIw&bvm=bv.89381419,d.d2s

It was for an FPGA implementation but it shows very clearly the arithmetic, logic and memory processes. In particular, have a look at the last part that shows hashrate against frequency for a single engine. Factor in the power requirement on a 28nm implementation and you'll see exactly why it's a bad idea right now.

Yes, I know Alpha has it's problems but I always respect any company that freely publishes knowledge for the benefit of others.

As for SFARDS I wish them good luck with their chip, especially with routing it.