The input-output relationship of innovation is usually classified as the research of innovation efficiency, which is of great significance to industrial development. Considering the lack of understanding of information industry innovation ability, this paper used data envelopment analysis (DEA) and the Malmquist index to study the innovation efficiency of next-generation information (NGI) industry in China by employing data from financial statements between 2012 and 2017. Partial least squares (PLS) regression model was used to explore the causal relationship between innovation input-output and between innovation output-lagged input of NGI industry, as well as possible mediating effects in these relationships from ownership structure, ownership concentratial clusters, company size, and innovation effectiveness. Empirical results showed that: (1) The innovation input, output, and efficiency of NGI companies varied significantly. (2) The innovation input, output, and efficiency of most NGI companies varied significantly across the years. (3) The innovation efficiency and efficiency change of NGI company have grown continuously. (4) No significant causal relationships between innovation input-output, or between innovation output-lagged input were found. (5) Company size and innovation effectiveness had significant mediating effects on the input-output causal relationships. Therefore, this paper suggested that the governments may enact NGI policies that focus on the innovation input-output and output-lagged input relationships, and may take into account the differentiated needs of the NGI enterprises based on the company size and innovation effectiveness.