Abstract Purpose Supplying off-grid facilities such as astronomical observatories with renewable energy-based systems (RES) instead of diesel generators can considerably reduce their environmental impact. However, RES require oversized capacities to counter intermittency and comply with reliability requirements, hence shifting the environmental impact from operation to construction phase. We assess whether 100% RES scenarios are favorable from an environmental point of view, and discuss the trade-offs in systems with backup fossil generators versus 100% renewable ones. Methods In this comparative life cycle assessment (LCA), we study various RES supply systems to power a new telescope in the Atacama desert, Chile. We compare six setups, including 100% RES scenarios, namely photovoltaics (PV) with batteries and hydrogen energy storage; high-renewable scenarios, with fossil fuel power generation next to RES and storage; and a system combining PV with diesel generation. We base system sizing on a techno-economical optimization for the start of operation in 2030. Foreground data stem from recent life cycle inventories of RES components and 2030 electricity mix assumptions of production places. We assess environmental impact in the categories climate change, mineral resource depletion and water use. Results and discussion We find that 100% RES and high-renewable scenarios result in emissions of 0.077-0.115kg CO2e/kWh supplied, compared to 0.917kg CO2e/kWh in the reference case with solely diesel generation. 100% RES scenarios have a lower CO2e impact than high-renewable scenarios. However, the latter lower the mineral resource depletion and water use by about 27% compared to 100% RES scenarios. Applying hybrid energy storage systems increases the water use impact, while reducing the mineral resource depletion. Conclusions None of the six energy systems we compared was clearly the best in all environmental impacts considered. Trade-offs must be taken when choosing an energy system to supply the prospective off-grid telescope in Chile. We find high-renewable systems with some fossil generation as the better option regarding power reliability, mineral resource depletion and water use, while inducing slightly higher greenhouse gas emissions than the 100% RES scenarios. As remote research facilities and off-grid settlements today are mainly supplied by fossil fuels, we expect to motivate more multifaceted decisions for implementing larger shares of RES for these areas. To advance the LCA community in the field of energy systems, we should strive to incorporate temporal and regional realities into our life cycle inventories. To ease the path for upcoming studies, we publish this work’s inventories as detailed activity level datasets.