Right, that's a bit old for what we know about I2a-Din. I'd expect I2a-Din to be as old as the J2, R1b-ht35, etc. in the region, not younger, as it appears to be. So I guess Thracians to have had a plurality of J2 (but admit I know less about the non-I haplogroups in the region).
The core area of the Thracians is still south of where I2a-Din apparently expanded from, then... unless we're thinking that they expanded northward, took I2a-Din peoples into their population, and then expanded southward again, but I don't see historical attestation of the second part of that equation. I suppose it's not impossible that there was genetic expansion within an existing population, it just seems like a poorer fit to the data.
To be clear, the high diversity area of I2a-Din-S is Central Europe, but the high diversity area of I2a-Din-N and therefore I2a-Din as a whole is Ukraine... S doesn't really have significant representation in Ukraine.
But again, I don't see why high diversity areas should be considered less than launching points of I2a-Din... and Central Europe and Ukraine were on the periphery of Thracian expansion, right? So the fit is backwards.
Of the populations you mention, I'm only particularly familiar with Cretans and Kurds. Kurds are data deficient and need more research... we only know the subclades of about 3 or 4 Haplogroup I Kurds, and they all seem to be primarily Central European type subclades (I2a2a/old I2b1, I1). Cretans are also somewhat data deficient, but we have an idea that their I2a-Din is low relative to mainland Greece and probably younger, and their other main clade is I2c-B-Z, which is actually probably even younger, and may have expanded within the Republic of Venice. I don't think either Kurds or Cretans have any particularly ancient I subclades that have been discovered so far.
Personally, I think that Southeast Europe doesn't have ancient I in the same way that Central Europe and West Europe do. See
the Paleolithic Remnants map and notice the dearth of subclades that have been shown to be ancient in the area.
It's doubtful that I2a-Din is even that old... likely, its coalescence age is younger, even if we find an outlier that pushes the TMRCA back.