Neural production networks: AI’s infrastructural geographies

Abstract

It is commonly argued that a handful of technology firms own the infrastructure that underpins the proliferation of artificial neural networks. But little is known about how this concentration of computational resources manifests itself in new geographies of production. To address this disparity, this article introduces the theoretical framework of neural production networks: geographically dispersed but computationally enveloped production arrangements powered by artificial neural networks. The article substantiates this framework by probing the role of Amazon, Google and Microsoft as lead firms in neural production networks. By reconsidering three key categories of production network scholarship – value, embeddedness and power – in light of those lead firms, the article opens a space for economic–geographical research on artificial neural networks.