Anyone who has developed training content - specifically hands-on, short format, data science training - will know that it is expensive and time consuming. As a rule of thumb, one can expect to spend 15-20 hours of development time per hour of delivery time. Even once the main work is done, there are still ongoing maintenance demands, from correcting the inevitable typos and adapting to software version updates, to major rewrites as best practice techniques evolve. Few Australians training organisations have a funding mandate for developing training material for general use. Instead most of us build training for our own communities in response to demand from those communities, allocating our limited resources to the areas of perceived highest demand. Notwithstanding this local focus, most training organisations operate in similar research environments with similar community demands, so there are real opportunities for benefit from collaboration between those organisations. In this session we will discuss how such collaboration could occur and whether a standardised national agreement around sharing of training material is an achievable outcome. This session will not discuss co-delivery opportunities, institution-specific content such as HPC training and data management, or consolidation of existing similar workshops into a single course. These are important topics, but we don���t have time to do them justice today.