To mitigate greenhouse gas (GHG) emissions, the importance of carbon capture and utilization (CCU) is continuously

increasing. Among alternative CCU options, electrochemical CO 2 reduction (CO2RR) is considered promising, as it can

benefi t from renewable electricity, which results in much less GHG emissions during the operation. Although CO2RR

remains at a low technology readiness level, its long-term deployment planning is worth an investigation, as the grid mix

is evolving to include more renewable energy resources. Conventionally, net present value (NPV) is used to evaluate the

economic viability of commercialization projects, where the future cash fl ows are discounted to determine the present values.

However, it cannot respond to dynamic uncertainties such as government policy and energy prices, because it makes

all the decisions at the current timestep. To deal with dynamic uncertainties, in this work, real options (such as expansion

and delay) are adopted to formulate a multi-period deployment planning problem for CO2RR. To solve such a problem effi -

ciently, a reinforcement learning-based algorithm is used. The results obtained by solving this problem are compared with

those obtained on the basis of NPV to highlight the necessity of multi-period deployment planning. Also, several scenario

studies are conducted to evaluate the impacts of diff erent uncertainties on the deployment of CO2RR with carbon monoxide

production as an illustrative example.