Whether the environmental benefits of emerging technologies are outweighed by the environmental impacts of producing and using scarce technology metals remains an open question. We present a three-level approach to assess how increasing environmental impacts on the resource provision level affect the overall impacts on the product level and on the service level, at an early stage of technology implementation. The approach is described based on a case example: we evaluate the environmental impacts of different supply options for lithium carbonate (Li2CO3) required for the production of Li-ion batteries and their influence on the environmental impacts associated with an EV (electric vehicle). We applied the methodology of LCA (Life Cycle Assessment) and considered the production of Li2CO3 from three different deposit types: natural brines (currently dominant), ores (less common) and seawater (hypothetical future option). For each of the three supply options, we established an inventory dataset for both favorable and unfavorable processing conditions. The inventory datasets were combined with those used in a recently published LCA, which compared the environmental impacts of an EV with those of an ICEV (internal combustion engine vehicle). The results of this study indicate that the environmental impacts of Li2CO3 production as a percentage of the total transportation impacts caused by an EV are currently negligible. Only if seawater was used under unfavorable processing conditions, these impacts could outweigh the environmental benefits of EV over an ICEV; however, the uncertainty is high due to the limited data availability regarding future lithium production processes. The break-even point for the environmental impacts of 1 km driven with an EV and with an ICEV would be reached only if the impacts per kilogram of Li2CO3 were increased by about two orders of magnitude (more than 200 times higher for the impact assessment method Cumulative Energy Demand, about 450 times higher for Global Warming Potential and about 100 times higher for ecoindicator 99).