Abstract: Crowd innovation spaces are important carrier for implementing the strategy of "Mass Entrepreneurship and Mass Innovation" in China. Clarifying the sources of loss in the innovation and entrepreneurship efficiency of crowd innovation spaces in China is of practical significance in promoting the high-quality development of crowd innovation spaces. Based on the operation mechanism of crowd innovation spaces, this paper first divides the innovation and entrepreneurship activity in crowd innovation spaces into investment stage and the incubation stage in tandem. Furthermore, the incubation stage is divided into two parallel sub-stages: the incubation sub-stage of entrepreneurial teams and the incubation sub-stage of enterprises. Then, taking 30 provinces and cities in China as research samples, this paper constructs a two-stage hybrid network DEA model to measure the innovation and entrepreneurship efficiency of crowd innovation spaces. Finally, the Meta-Frontier theory is used to locate the sources of efficiency loss at two levels: technology frontier inefficiency and management inefficiency. The results show that: (1) Under the meta-frontier, the national average innovation and entrepreneurship efficiency of crowd innovation spaces in China is 0.724 during the study period, and more than 50% of the regions have an efficiency below 0.75, so there is still much room for improving the innovation and entrepreneurship efficiency of crowd innovation spaces in China. In terms of stages, the average efficiency values of the investment stage, the incubation stage of entrepreneurial teams and the incubation stage of enterprises are 0.856, 0.819 and 0.828, respectively. The efficiency under the meta-frontier does not exceed the corresponding efficiency under the group frontier. (2) There are significant differences in the distribution of technology gap ratio in the investment stage and the incubation stage of entrepreneurial teams in the three regions of China, but there is no significant difference in the distribution of business incubation stage. The technological frontier of the eastern region in the stage of investment is closest to the national technological frontier. The central region has the best technological frontier in the incubation stage. And the technology gap ratio of the western region in each stage is lower than that of the eastern and central regions. (3) About 16.7% of the samples belong to the ineffective type of high technology and high management, including Anhui, Fujian, Inner Mongolia, Xinjiang and Gansu, which need to be paid special attention to in the development of crowd innovation spaces in the future. About 53.3% of the regions have more efficiency losses due to differences in technological frontiers (such as Guangxi, Yunnan and Chongqing) or ineffective management (such as Shandong, Zhejiang and Jilin). All regions need to formulate targeted optimization schemes according to the sources of efficiency loss in different stages. In general, in the future, the eastern region should focus on optimizing the internal governance structure of the crowd innovation spaces to reduce endogenous barriers to efficiency enhancement, such as strengthening the supervision and management of the whole process of government financial subsidies. The western region should strive to optimize the regional innovation and entrepreneurship environment, accelerate the narrowing of the regional technology gap, reduce the exogenous resistance to efficiency enhancement, and gradually improve the management level. The central region should give full play to its location advantage and strengthen the coordinated development with the eastern region.