What's been said is that they are sure they will get a 30% price reduction (pr. kWh) on the cells from the GF. JB have said that he hopes for - and see it as a possibility - to get up to a 50% price reduction on the cells when the GF is in full production. I'm unsure if they have calculated in a price reduction from new/better chemistry or other development on the batteries from the planing in 2013/14 until full production as planned for 2020.
Tesla is very much run by engineers. They wouldn't be speculating on a 30% price reduction unless they had some pretty good ideas of where the savings would be seen. For one thing, they are drastically shortening the logistic trail. Right now the lithium and other materials for batteries is mined who knows where, then transported (most likely by ship) to somewhere in Asia where the batteries are made, then the finished batteries are shipped to the US on a ship. Along the way, they pass through many hands and several ports. The various components are probably made in different places and there is more shipping to get all the pieces together in one place for final assembly.
With the Gigafactory, the raw materials are being sourced in North America where they will likely be put on trains and shipped to Nevada. There everything is processed and manufactured into batteries all in one place. When done, they are put on another train and sent a fairly short distance to Fremont. That is a savings right there that is probably 10-20%. They might have some plans for better battery layouts than what they are currently using. A professor at MIT has come up with a manufacturing process for Li-Ion batteries that reduces the time from 24 hours for making cells to about 10 minutes (the method in widespread use today needs to wait many hours for things to cure before final assembly). I don't know if Tesla has approached him for the rights, or if they reverse engineered it and came up with a similar method that got around the patents. That could be more than 30% savings there.
That's just a couple of cost savings that came to me off the top of my head. I'm sure Tesla has a team working on battery technology and how to make them much faster and easier. Considering how much innovation they have come up with in other areas, I wouldn't be surprised if they haven't come up with some new techniques the world doesn't know about. Probably the hottest area in electronics today is research into new batteries. Research is ongoing into both new battery chemistries as well as improved manufacturing techniques for existing chemistries. I suspect Tesla is more focused on the latter, but there have been quite a few academic papers presented on new techniques for Li-Ion batteries and I'm sure the lab at Tesla has been focused on how to manufacture them in bulk in the best way for reliability, longevity, and price.
I noted when the 90KWh battery was unveiled, Elon said it was necessary for the P90DL and it was going to be standard on the Model X, but he didn't think it was worth upgrading 85KWh batteries. Reading between the lines, I think he was hinting that there was going to be some kind of battery breakthrough once the Gigafactory was up and running.
Long term, I expect Tesla will switch to the next battery chemistry that proves itself better than Li-Ion. The big winner will be when someone comes up with a battery technology that recharges well without degradation and has the same energy density as gasoline or better. Right now Li-Ion batteries have about 1/30 the energy density of gasoline. If the Model S had a battery with the energy density of gasoline, it would get around 9000 miles on a charge. There is nothing even close to that kind of energy density today though.