Hydrogen fuel cell vehicles (FCVs) are one of the cleaner technologies that can be used to mitigate GHG emissions and local air pollution from transport sector, but they are currently too expensive to compete with dirtier alternatives. One important mechanism through which costs decline is learning-by-doing, the process by which cumulative global deployments lead to cost reduction. This is a highly uncertain but critically important aspect of the future for hydrogen technologies. This paper analyses the cost effectiveness of FCVs in meeting global climate targets by implementing learning-by-doing endogenously in the TIAM-UCL global energy system model. FCVs share many components with other electric vehicles. The paper examines the complementarity and competition between FCVs and other electric vehicle types by applying endogenous learning to three key technologies: fuel-cells, batteries and electric drivetrains, which are shared across different vehicle types, including electric and plug-in hybrid cars and fuel cell buses. Discussion of results will focus on uptake of fuel cell vehicles in different long-term decarbonisation scenarios, and will highlight the rates at which technological progress must take place if FCVs are to succeed.