Many of the environmental impacts associated with hydraulic fracturing of unconventional gas wells are tied to the large volumes of water that such operations require. Efforts to develop nonaqueous alternatives have focused on carbon dioxide as a tunable working fluid even though the full environmental and production impacts of a switch away from water have yet to be quantified. Here we report on a life cycle analysis of using either water or CO for gas production in the Marcellus shale. The results show that CO-based fluids, as currently conceived, could reduce greenhouse gas emissions by 400% (with sequestration credit) and water consumption by 80% when compared to conventional water-based fluids. These benefits are offset by a 44% increase in net energy use when compared to slickwater fracturing as well as logistical barriers resulting from the need to move and store large volumes of CO. Scenario analyses explore the outlook for CO, which under best-case conditions could eventually reduce life cycle energy, water, and greenhouse gas (GHG) burdens associated with fracturing. To achieve these benefits, it will be necessary to reduce CO sourcing and transport burdens and to realize opportunities for improved energy recovery, averted water quality impacts, and carbon storage.