Cost-efficiency is important. However, other things are important too.
Having something earlier is worth more than having something later. Earlier launch capacity means constellations can reach a degree of operational status earlier. The value of constellation size is usually not linear. In some cases marginal returns will follow an inverted U-shape, and most cases will have strictly diminishing returns. Having a constellation become operational earlier means that the industry can start earlier to accumulate the experience of fielding and operating such a constellation, and the same is true of downstream applications. As I understand it, under ITU rules, as well as under the rules of many countries, spectrum rights are allocated on a first come first served basis, which would be relevant in the areas where there are spectrum conflicts with American competitors.
Schedule certainty has great value all on its own. Surprise delays in launch causes cascading delays to payloads and the downstream applications of those payloads, costing money. If every Chinese launch company focuses on chasing the latest and greatest in launch technology despite the schedule risks that entails, it could lead to a scenario where a lot of payloads remain grounded for years, causing payload companies and the people who intended to use the payloads to lose enormous amounts of money. Delays to launchers are not independent events when the launchers are highly similar. So the fact that there are several development projects running in parallel does not remove uncertainty. If there is too much launch capacity uncertainty, many companies will delay investments in payloads and downstream applications, because of the risk of losing a lot of money on delays. It could mean that China encounters a situation of having a lot of launch capacity and not enough payloads to launch.
Many other Chinese companies are working on rockets that are potentially more cost-effective than CZ-5-8, but will take time to finish and suffer from schedule unpredictability. So it makes sense if CASC decided to prioritize bringing balance the industry by focusing on rockets that are less cost-efficient but can be had soon and on a highly predictable schedule.
Some more thoughts on the CZ-10A, why I think a very large fleet might not be built, and why it might not necessarily be advantageous to do so:
Suppose first launch of CZ-10A in 2026. It took SpaceX two years to successfully land a booster from beginning to try in 2013. Assuming CASC follows the same path, it means CZ-10A boosters won't start accumulating before 2028.
The state of China's launch industry in the early 2030s will be very different from now. There will be a lot more rockets operational, providing a solid baseline launch capacity that won't rest on CASC alone. Thus, by then, it makes sense for CASC to focus on the cutting edge at the expense of the established, and they might start converting some YF-100 production capacity to YF-215 beginning around 2031. That way they can hit the ground running when the CZ-9 is ready to launch in 2033, like SpaceX which started large scale production of Raptor before SS/SH launched and before Raptor itself had finished development.
The CZ-10 project will start to consume engines from 2027 onward. A first launch in 2027 will consume 29 YF-100. Perhaps an unmanned lunar landing rehearsal in 2028 will consume 58 YF-100. From 2029 onward, it will consume 116 YF-100 per year, assuming two manned landings per year, like Apollo. I think this could constitute a large chunk of total YF-100 production at the time. I think it will be hard to dramatically scale up YF-100 production, because competition for relevantly skilled labor will be ferocious. There will be competition with other projects within AALPT (6). There will be competition with turboshaft and turbofan engines, such as WS-20 and CJ-1000. There will also be competition with other Chinese launch companies, some of which have grand plans to produce 300-500 engines per year (3, 4, 5). And it would make sense to fund those plans, because they likely have better rockets than CZ-10A.
It is interesting to compare CZ-10A with Falcon 9 FT. F9FT can send 14,460kg to 51.6 degree inclination 400km circular LEO (ISS orbit) using ASDS landing (1), with launch mass 549t. That gives a payload mass fraction of 2.63%. CZ-10A's payload mass figure for LEO can't be referring to a worse LEO orbit than that. CZ-10A's barge landing payload is 14t, with a launch mass of 750t. This gives a payload mass fraction of 1.87%. So F9FT has a payload mass fraction that is 41% higher. This is despite CZ-10A benefiting from scale, staged combustion, and the elimination of landing legs. There seems to be something very suboptimal with the design. The choice to eliminate landing legs and use unproven wire-catching instead, while innovative, raises the suspicion that 14t is indeed the absolute limit, and that CZ-10A wouldn't be able to send the new crewed spacecraft to the space station if it had conventional landing legs.
The apparent readiness to employ suboptimal and expedient kludge solutions for the rocket, makes me doubt how reusable the engine will be made. Using an oxidizer-rich preburner, for an engine that wasn't originally designed to be reusable, does not intuitively seem conducive to a long and problem-free engine life, unless it is thoroughly redesigned. Which might be a bad idea, if it means taking materials R&D resources away from the YF-215.
I don't see why an expendable CZ-10A would bring big cost savings. As an expendable rocket, it consumes 8 YF-100 to send 18t to LEO. This compares with ~25t to LEO using 8 YF-100 and 2 YF-77 for CZ-5B. CZ-10A requires only one fuel type, one engine type, and one tank diameter, however the CZ-5 launch pad and production lines for the YF-77 and 3.35m tanks already exist anyway. The YF-77 production line can't be shut down anytime soon, because the CZ-10A doesn't replace the CZ-5 for high orbits and deep space missions.
CZ-10A is a CALT rocket. Historically, about 40% of CASC rockets launched have been SAST models (2), and I assume there will be a similar ratio in the future. SAST rockets will need engines. As far as I know, SAST does not have the facilities to make 5m diameter propellant tanks for the CZ-10A, nor does it have a practice of adopting CALT rocket models. Maybe SAST will eventually develop a rocket that uses another engine, such as YF-209. Though, a production line building YF-209 is a production line not building YF-100, so that would impact YF-100 availability for CZ-10A too.
Thus I do not think there will or should be a large fleet of CZ-10A, the exception being if YF-215 development suffers severe delays.
I will add that I don't want to pretend to preclude the possibility that I am wrong, in part or whole. I just think that there are reasonable explanations for why CASC would chose to prioritize CZ-5-8 over CZ-10A, that don't involve stupidity or organizational paralysis, and that one should not measure the state of CASC merely by how fast CZ-10A is rolled out.
(1)
(2)
(3)
(4)
(5)
(6)
