CAC is able to produce hundreds of planes, even J-20s per year. As any plane maker with similar experience and position would be.
Ramping up to such levels of course takes time and money. So many people in the whole production chain to hire, train, etc. Expansion of whole infrastructure and said production chain. Whether it'd take a few years or a decade is besides the point.
But in reality that's not the situation today nor does it seem it will happen in the next decade or so. The only situation that'd spur such a change would be if the Chinese government said - "New World War is not only inevitable, but it's going to happen within the next X years. We're million percent sure of it". And then if said government went to the PLA and asked them "what would it take to win said war?" And, among other things, the PLA said "we'll, we'd need to double the size of our air force, and have it made up mostly of 5th gen planes. But it needs to be done quickly". So then the PLAAF would suddenly get obscene amounts of money, and keep getting them for years.
In such a situation we'd likely see those Defense budgets skyrocket from a 1.4 trillion Yuan to two trillion Yuan per year. (roughly speaking).
And then within several years we'd indeed see massive numbers of new units getting J20s, whole new units standing up and so on.
Anyway. We're not seeing that because it's not happening. Because PLAAF did not get said amount of money. Because the government doesn't see the war as an absolutely unavoidable and/or doesn't see it happening for sure within the next decade.
But as for the capability of Chinese planemakers to do that once and if such a decision comes - said capability is there. It's "only" constrained by time and money.
I would not be so quick to make such big conclusions.
It’s precisely because it takes years to spin up additional production capacity from start to delivery that our ability to observe such developments are understandably lumpy.
We won’t see all the incremental, behind the scenes hard work until the fruits of said work and investment start getting delivered to frontline units. In fact probably not even then as our ability to track new deliveries are massively limited to what China chooses to let us see and what slips through the net of its internet censors.
I think people may be surprised at the J20 production rates, as I think work to boost said production rates would have started years ago, and the first signs of that feeding through into additional production would have been around the time China started cracking down on wall climbers and the effective live updates from CAC abruptly ended.
I think that, and the preceding years of tolerance to such detailed reporting was very deliberate and well done effort to conceal the true annual production rates in the subsequent years.
By allowing a pretty consistent and relatively complete dataset to be collected through such unofficial leaks in the previous years, and then ending the leaks, the most logical course of action for outside observers would be to simply extend the original datapoints by time to arrive at the expected current total production numbers.
Now imagine if that original dataset was allowed to be collected precisely because CAC was only spinning up production rates and maybe even slowing down production as they trained people up for the new line(s) to replace the J10 lines and/or as they re-rolled for engine changes etc.
So the observed production rate might be 24 planes per year for example, but the true production rate might be 36 or even 48. This can make a huge difference if you are using the existing line production rate to estimate the new total production capacity of CAC once the J10 lines have finished converting to J20 production.
It’s easy to make such educated guesses, but nearly impossible to substantiate and prove it due to the lack of solid datapoints on current and future J20 production rates. More so because of the strength of the original dataset means that many will demand equal or better data before they will accept any assumed peak production rate per line.