Regenerative design replaces the linear material flow that characterizes industrial economies with cyclical flows that maintain and restore the living systems economic activity depends on. A degenerative economy extracts raw materials from the Earth, converts them into products, uses them briefly, and discards them as waste. A regenerative economy designs its material flows in loops — restoring soils, replenishing water tables, maintaining biodiversity — such that economic activity contributes to ecosystem health rather than depleting it.
Applied to AI, regenerative design has two registers. The first concerns infrastructure: the data centers, devices, and semiconductor supply chains that currently operate on a linear take-make-use-lose basis that is ecologically destructive at every stage. Regenerative AI infrastructure would require circular material flows — device recycling at scale, data center design for reuse, renewable energy procurement coupled with ecosystem restoration.
The second register is more suggestive. The human activity that AI enables — the ascending friction that shifts workers from mechanical execution to judgment and care — has inherently regenerative characteristics. Care work builds social capital without consuming material resources. Judgment improves decision quality without increasing throughput. Creative direction channels productive capacity toward meeting needs rather than generating waste.
The regenerative potential of ascending friction is conditional. It depends entirely on where the freed human energy goes. The Berkeley study documented where it currently goes: into more work, more tasks, longer hours, blurred boundaries between labor and rest. The freed capacity is captured by the growth logic and converted into additional throughput — more production, more output, more stuff. The ascending friction generates a surplus of human cognitive capacity, and the economy consumes the surplus before it can be invested in regenerative activities.
A doughnut-compatible direction of AI-freed capacity would build care infrastructure, ecological stewardship programs that employ human judgment in the restoration of living systems, and working-time legislation that connects productivity gains to reduced hours rather than increased output. These are institutional design choices, not technological ones. The technology supports either outcome equally. The economic logic determines which prevails.
Raworth draws on the circular economy work of the Ellen MacArthur Foundation, the regenerative agriculture lineage running through Wes Jackson and Allan Savory, and biomimicry scholarship (Janine Benyus). Her synthesis extends these approaches from specific sectors to the whole economy.
Linear to cyclical. The fundamental move is from take-make-use-lose to closed-loop material flows.
Infrastructure and human activity. Both registers matter: AI's physical infrastructure needs regenerative redesign, and the human capacity AI frees needs regenerative direction.
Conditional potential. Regenerative outcomes depend on institutional design; without it, freed capacity is captured for more throughput.
Care, judgment, and stewardship. The regeneratively-aligned uses of AI-freed human time.