In recent years, corporate actors have been actively shaping the normative debates, public policy and regulatory infrastructure related to AI governance. In this paper, Yishu Mao, Vanessa Richter and Christian Kaztenbach show how they achieve this by analysing corporate communication materials in China, Germany and the US from 2017 to 2023. Drawing on the concept of sociotechnical imaginaries, the analysis shows that companies in the three countries have strategically constructed imaginaries such as ‘Responsible AI’, ‘Trustworthy AI’ and ‘AI made in Europe’ through communication, network formation, and institution building. These imaginaries both respond to regulatory pressures and shape public policies that ultimately serve corporate interests. Although companies construct situated imaginaries that draw on local socio-political resources to respond to national policy agendas and public sentiments, these imaginaries were leveraged to advance policy recommendations that lend themselves power in building future AI governance infrastructure and institutionalisation. This analysis demonstrates how globalised corporate visions of innovation negotiate political tensions and socio-cultural differences for their dominance, and highlights the need for critical reflection on the implications of concepts such as ‘Responsible AI’, ‘Trustworthy AI’ and ‘AI for Good’ for public imagination and future trajectories of AI governance.