The 115-year history of the Hall-Heroult process for production of aluminium has been characterised by a sustained technology-driven improvement in costs and productivity. In the past decade the industry has effectively transferred best practice into older technology with improvements through computer control of the process, and redesign of cell linings using computer modelling . Incremental development is now at the point where diminishing returns will apply for further technical improvement effort however, until the gap between technical limits and current best practice is addressed through major process redesign. The process remains a long way short of its technical limits in energy consumption, carbon anode consumption, cell life and product purity. Investment cost in new capacity is very high, supporting the retention of older and less efficient technology and hence limiting the application of best practice in environmental performance and occupational health exposure. The industry is a major emitter of greenhouse gases yet the process has the potential to release oxygen as the primary waste. All of the above issues will be addressed in the new century, but only through commitment to the necessary long-term research and development that is required to achieve a redesign of the process.