Abstract: Decentralized Finance (DeFi) is often conceptualized as unregulated, disintermediated and separated from legacy institutions, and pertaining to cryptoassets. This paper, focusing on the case studies of JP Morgan’s FedSyn and WeBank’s FL-Market, broadens the definition to include forms of decentralized financial computation and the decentralized ownership and trading of artificial intelligence data. Furthermore, it invites to include experiments championed by established financial institutions. In this way, more subtle stakes of DeFi come sharply into focus: first, at the level of infrastructure, distributed machine learning computation performs an ambiguous politics of decentralization over cloud computing, countering tendencies towards agglomeration while potentially reinforcing the power of established through “on premise” datacenters. Second, at the level of data, the combination of distributed datasets and of synthetic data produces tensions between data rents and data assetization: data are no longer scarce, yet they are more immediately monetizable, resembling the explosion of financial derivatives. Third, changes in the political economies of both infrastructures and data have momentous consequences on users, whose behaviour is now measured against a yardstick of synthetic and distributed “dividuals” produced by generative algorithms trained on distributed datasets and infrastructures. Financial geographies should then revisit some of the literature on financial subjectivities to see how new subjects are summoned through generative AI algorithms.

This session will be delivered via Zoom. The joining link will be sent to you, by email, 3 days prior to the session.