I have asked for empirical data if there is any to back up the claims. If there is none, then they remain assertions.
Of course these are calculated assertions. You are asking for the kind of data that is classified from both sides as proof, which is impossible. But something like volume amortization happens to be a fundamental basic of the electronics industry.
Let's say it took you $100 million for R/D to develop this chip. To produce one chip, along with hardware, manufacturing cost and corporate costs, this single chip would be upward of $100 billion, since this entire single chip has to pay for the R/D cost. Build ten of these chips, the R/D cost is divided into ten, for $10 million each. A hundred chips puts the cost down to $1 million each. 10,000 of these chips reduces this cost to $10,000 each, and 100,000 of these chips reduces the cost to $1000 each. At some point the R/D cost is reduced so the greater fraction of the unit cost becomes material, manufacturing, marketing, shipping and administrative cost, all of which are also volume amortized.
So let's say, we have an unknown, an X cost, in the R/D developing the T/R module for the Type 346. As I mentioned, there are over 5,000 T/Rs per face, and four faces that makes it over 20,000 T/Rs. But as I also mentioned, they are using a QTRM, so four T/Rs per module, and that means around 5,000 modules per ship. Then you multiply that with every 052C/D that is now in service, being fitted and produced, + every 055 being produced, both now and the future. The more of these ships are made, the more of these radars are made, the cost curve for each module drops and so does the cost of the radars. Even if later ships have improved modules from the original design, the R/D cost for these improvements would be small and minor compared when designing the device itself from scratch. You can further amortize the cost of these modules if they are used on the 052E and 054B for instance. Now compare this scale of production with those for any other navy in the planet.
What's the cost of the R/D in China? We don't know exactly but the cost of salaries there are still much cheaper than in the US. With the IP copying that the US is so furious about on China, the R/D costs are even lower and less licensing costs are also paid. With China having its own foundries, being the dominant global producer of refined Gallium and Yttrium, there goes your materials and manufacturing costs. For all other human related costs, that's going to be reflected from the cost of living --- rental and housing expenses, cost of power and water, food, medical, etc,. All of which are also lower in China than the US but both are also rising. The fixed costs will be amortized with higher production, which also in turn allows manufacturing costs to lower.
It might be said that its possible that the X-band set on the mast of the 055 might be more expensive than the main set of 346B, because it is a brand new design, and its hard to say how much R/D is shared with similar X-band modules used on the J-20, J-10C, J-16, and J-11D radars. The costs of these radars are going to be front loaded until enough of them are built to amortize the cost.