PESA only has one emitter to feed the entire array even though the array consists of elements each with its own phase shifter, whereas AESA has one emitter inside every element. AESA wins by sheer redundancy. AESA has the advantage of being able to send in multiple frequencies by grouping sets of elements to to one frequency, but I won't call this advantage to be that important since you can always time share different frequencies with a single emitter.
AESA actually works better for China's technology and industrial infrastructure. Powerful PESAs need powerful emitters to match, something the Russians have developed very well. For all its worth, its actually easier to develop an AESA because its better to just use lots of small emitters instead of one godzilla emitter. China skipping into AESA may have something to do with this. It works with the current industrial infrastructure, since AESA elements currently use GaAs, a semiconductor used mostly with LCD displays, and guess who makes a lot of LCD displays.
The arrays on the KJ-200 (not the KJ-2000), and the 052C are more likely to be AESAs too. Another AESA application is the SLC-2 artillery spotting radar.