The NSF has supported early forms of scientific cyberinfrastructure from the 1960s. Since about 2000, however, new cyberinfrastructure (CI) initiatives have gathered momentum, guided by an increasingly comprehensive vision of CI as a principal agent of change for a new era of large-scale, distributed, data-intensive, collaborative science in virtually every domain. EarthCube, as a large-scale international CI project provides a fertile ground for the observation of the negotiation and policy work necessary to facilitate distant collaboration and the accumulation, provision and sharing of scientific knowledge and resources (data, tools, models). As a collaborative design process drawing from both the policy and scientific worlds, EarthCube is a site where the complex relationships and negotiations between governance and particular visions of novel science are particularly evident as well as being a microcosm of the co-constructive processes that produce relevant knowledge for both policy and science. This writing explores the relationship of policy work and the production of scientific knowledge, and provides an account of the complex interactions between regulatory, legislative, and organizational work with knowledge production practices in the geosciences through the lens of research funding and cyberinfrastructure development both historically and ethnographically.