Societal advances have been market by human's ability to understand and engineer materials from the stone, iron and bronze eras to contemporary dominance of silicon and plastics. Now materials scientists frame their discipline as building atom by atom to engineer the properties needed at every length scale. This capability in computational materials science plays a critical role in designing materials for essential sustainability advances in catalysis, energy storage and decarbonizing industrial production. However, the energetic and environmental burden of computing also comes with significant computational cost. My presentation will cover two topics, first a set of cases of how we've leveraged natural language in scientific texts to uncover synthesis and manufacturing insight within the materials science domain as well as efforts to understand the climate implications computing activities across the supply chain.