What's interesting is days destroyed is basically at zero. All the coins being moved to exchanges to be sold are probably very new.
As I noted before,
(1) the BTC volume
Vex transferred to/from the exchanges is almost certainly a very small fraction of the total transaction output volume
Vtot in the blockchain (600'000 BTC/day now). Therefore, even if
Vex includes a few thousand old coins per day, it will hardly affect the plot of bitcoin-days destroyed (2 to 25 million BTC-day per day).
(2) For the "bitcoin-days destroyed" formula, the age of the coin is when it was last moved, not when it was bought. So if someone bought 10'000 coins in 2010, moved 2'000 of them to a hot wallet two months ago, and sent these to an exchange today, there would be a lot bitcoin-days destroyed two months ago, but only 120'000 today, which will not show in the graph.
By the way, I didn't see an aswer to my question about the change-back outputs: are they excluded when the "bitcoin-days destroyed" is computed?
(I also don't know the guys how blockchain.info identify the change-back outputs to produce the
Estimated Transaction Volume chart. I tried asking in their "help" thread, but it seems that the person who created the charts was eaten by an alien giant ameboid blob, and they haven't noticed his disappearance yet.)