Controlled-Environment Agriculture (CEA)—a form of agriculture encompassing a spectrum of technological assemblages, such as hoophouses, greenhouses, and vertical farms—has been claimed by its proponents to represent a revolutionary vision for agrifood systems. The arguments for continued innovation and scaling up in CEA include not only efficiently increasing yields but also claims to progress in environmental sustainability, renewable energy infrastructure, local food production, job creation, and profitability.

In Kentucky, such claims were used to justify the founding of an agritech startup called AppHarvest which accumulated over $800 million in seed funding, venture capital investment, and loans from banks and other organizations (Adams 2024). Before going bankrupt in 2023, AppHarvest constructed four large-scale greenhouses—two at 60 acres, one at 30 acres, and one at 15 acres—in Appalachia which were subsequently auctioned off and are now operated by different companies. Despite AppHarvest’s failure, an effort toward fleshing out an “agritech ecosystem” in the state persists in the cross-sectoral consortium called Let’s Grow Together, which uses the Dutch Triple Helix model of collaboration to unite industrial, research, and government partners from Kentucky and the Netherlands.

Using Let’s Grow Together as a case study and entry point, I have paired ethnographic methods and semi-structured interviews with key stakeholders in CEA from Kentucky and the Netherlands to interrogate how CEA, as a field, operates. My research question, which has been shaped iteratively by the research process, is: What are the visions offered by agents within CEA and how are they enacted? I use the term “vision” to allow for a dialectical analysis of the top-down, model-driven vision of control which is central to how CEA operates as a system with the simultaneous vision of care displayed by agents’ understanding their own practices within the field. This analysis also allows for the mode of vision embedded in high modernist projects, “seeing like a state” (Scott 1998), to be linked to a phenomenological vision which is operative at the level of an individual agent’s body or “habitus” (Bourdieu and Wacquant 1992).