There is a critical need to incorporate life cycle assessment (LCA) into research and development of renewable energy technologies, so that cradle-to-grave environmental concerns are identified early and communicated to technology developers. For example, environmental analyses of photovoltaic (PV) technologies may call attention to energetically burdensome processes with room for improvement, and can be used to compare the net energy balance of competing PV technologies. However, existing LCA frameworks are largely retrospective (i.e., requiring detailed data from existing industries and supply chains), and are thereby unable to provide timely information to decision makers. Large growth in the photovoltaic industry necessitates the development of anticipatory LCA methods, which can be used to explore potential environmental impacts of technologies and industries as they evolve. While the economic experience curves (i.e., $/watt) in PV project the costs to continuously come down as efficiency improves, it is not obvious that environmental experience curves (i.e., embodied energy/watt) are also monotonically declining. We review the boundaries and assumptions of all published PV-LCAs, and develop environmental experience curves for mono-Si, multi-Si, amorphous Si, and cadmium tellurium PV cells. The curves show decreasing manufacturing energy burden for silicon cells and relatively little improvement for thin film technologies. Using these environmental experience curves, we calculate the Energy Returned on Energy Invested (EROI) from cradle-to-use for each technology, plotting historic improvements for each technology and calling attention to the different rates of improvement. Results suggest alternate research and policy agenda, for example, silicon research should focus on decreasing manufacturing and supply chain investments while the net energy production of thin film PV may be improved largely through increases in efficiency.