simplest answer is that Alibaba and Baidu have more party members as stakeholders

but its not likely that any uncontrollable LLM can start spitting out accuracy or things unhelpful to Beijing’s ethos there and be allowed to operate

the model or the service filtering the model has to be controlled