Innovation uncertainty shapes divergent pathways toward decarbonization
by Theo Moers, Davide Fioriti, and Gernot Wagner
Abstract:
Clean-energy technologies are getting better and cheaper, but just how much is uncertain. Falling costs for solar, wind, and batteries have repeatedly changed expectations for the power-sector transition. Yet most models still use exogenous cost scenarios or fixed learning curves, limiting feedbacks among deployment, innovation, and system design. We here examine how modeled innovation uncertainty is linked to clean-energy deployment pathways with no pre-defined carbon constraints or price. By coupling empirically validated stochastic learning processes for solar, wind, and batteries to a recursive global energy-system model, we trace how uncertain future costs and deployment feedbacks shape power-system pathways. The median 2050 system reaches ≥75% renewable electricity (10–90% range: ∼67–82%), with a median fossil electricity share of ≤20% (∼13–27%). Yet specific pathways and cumulative emissions vary widely. Pathways with earlier clean deployment lower emissions at little energy-system-cost premium, suggesting that policy evaluation should go well beyond expected costs and instead focus on early investment and transition pathways.
Draft: "Innovation uncertainty shapes divergent pathways toward decarbonization" (18 September 2026); preprint available upon request.
Presentation: SURED 2026 keynote (1 June 2026).
