<p>The study of Chinese labour politics has returned to the centre of scholarly interest as China has increasingly become involved in global production and trade. As the incidence of labour dispute and workers’ strikes continued to soar, ubiquitous cases of labour rights abuse have been widely reported by international media and academics. The literature of Marxist international political economy has long predicted the insurgency of the Chinese working class resulting from rising inequality, global capital movement and labour division. In contrast, traditional Chinese labour studies are inconclusive as to whether the Chinese working class has gained enough class consciousness to become a cohesive agent for social and political change. This research examines how rising economic, social and political inequalities have impacted on the Chinese working class’s agency. The research shifts the focus from top-down structural analysis to workers’ agency itself, with an emphasis on their cognitive strength. The research was undertaken via a two-case comparative study of the Chinese working class in four megacities and four smaller cities. Data came mostly from statistics and field interviews. This two-case comparative study concludes that, overall, the Chinese working class had a weak behavioural strength, as manifested by inconsistent wildcat-style strikes, which had no clear political strategies. This research also concludes that the working class’s cognitive agency is weak and conservative, as manifested by a weak class identification, their poor understanding of democracy, their low willingness to participate in collective action, and their weak sense of class solidarity. I argue that inequalities and capital movement do not have a simple and unidirectional relationship with the working class’s collective agency. On the one hand, inequalities and capital movement can arouse the working class’s behavioural strength quickly. On the other hand, workers’ cognitive strength is more inert and does not correspond neatly to these two factors. The research findings show that the megacities are more economically developed, with higher inequalities, but with considerably weaker and more conservative working class agency; whereas the smaller cities are less economically developed, with lower inequalities, but with less weak and conservative working class agency. The addition of cognitive strength as a new dimension of working class study provides a pluralist analytical framework for the study of Chinese labour. The new Chinese working class are better educated and more individualised with three main characteristics - occupation-based, precarious, and conservative - which distinguish them from the older generations of workers who had a clear group identification, such as the SOE workers in the 1990s, and the rural migrant workers in the 2000s. These theoretical and empirical findings open up possibilities of new strategies for effective labour organisation that should be considered by labour NGOs, civil society and the government. These players not only need to manage the working class action carefully, but also need to better understand the workers’ complex cognitive situations.</p>