It would be too invasive on our part to write to another software system’s store without explicit instruction and use of a contract, which is what we do for our Maven plugin’s install task.
Also, the physical layout of Maven’s repository would not support the characteristics and features of our cache.
In summary, there are several reasons why using maven’s cache would be a bad thing for Gradle and one good reason (disk space optimization) and one debatable (state sync between the two tools). It’s just not going to happen.