Many low-carbon scenarios rely on carbon dioxide removal (CDR) to meet decarbonization goals. The feasibility of large-scale CDR deployment is highly uncertain, and existing scenarios have been criticized for overreliance on CDR. We conduct an expert survey on the feasible potential for CDR via bioenergy with carbon capture and storage, direct air capture and afforestation. We use the survey results to represent uncertainty in future CDR availability and explore the implications in an integrated assessment model. Stochastic optimization demonstrates that uncertainty in future CDR availability provides a strong rationale to increase near-term rates of decarbonization. In scenarios with high CDR uncertainty, emissions are reduced by an additional 10 GtCO2e in 2030 compared with scenarios with no consideration of CDR uncertainty. This highlights the urgent need to increase ambition contained in nationally determined contributions (NDCs) for 2030, to get the world on track to deliver 1.5°C and to hedge against an uncertain future CDR potential.