The University of Newcastle announced trials of two chemistry robots on 24 September 2026 as part of a six-platform research network across New South Wales. The university says the systems can test large numbers of chemical combinations and use machine learning to help select promising experiments. The immediate development is additional research capacity, with commercial benefits still to be demonstrated.
The Newcastle systems are installed at its Callaghan campus. They support medicinal chemistry and particle–liquid formulations, areas relevant to pharmaceuticals, mineral processing and water treatment. Newcastle says the robots can work around the clock, extending the number of experiments researchers can explore. It has not reported a resulting medicine, commercial water-treatment process or quantified industry productivity gain in this announcement.
The wider network connects Newcastle with three other NSW universities. Wollongong, UNSW and the University of New England are also involved. Newcastle identifies an A$1.99 million Australian Research Council infrastructure grant supporting the collaboration. In its release, Professor Erica Wanless describes the potential to investigate formulations more systematically and give students experience with automation and artificial intelligence.
Wollongong’s laboratory overview provides a more concrete picture of the experimental work. Its part of the capability includes robots that dispense materials, prepare samples and screen reactions. It describes optimisation across variables such as ingredients and process conditions, with objectives including yield, selectivity and physical properties. These are laboratory operations that can be repeated and measured.
The same overview puts the broader investment at A$2.7 million. That total includes the ARC and a consortium of NSW universities, whereas Newcastle’s A$1.99 million figure refers to the grant. The distinction avoids understating the contribution needed to create and operate shared research infrastructure. Wollongong describes six platforms across three sites and a remote node in Armidale.
Automation and machine learning play different roles within that process. A robot can carry out a set of preparation and testing steps, while a model can help prioritise which combination should be examined next. That is the logic of the capability described by the universities. It is not a guarantee that the most promising laboratory result will remain effective when production conditions change.
A practical evaluation should therefore look beyond the number of experiments. Useful evidence would include reproducible results, the resources needed to obtain them and whether a candidate formulation performs under the conditions of its intended use. Those are questions for the network’s subsequent research, rather than outcomes established by installing the equipment.
For Australian researchers and businesses, shared access could make more systematic testing possible. The network’s value will depend on turning experimental capacity into findings that other scientists can validate and industry can assess. Newcastle’s robots make that work more feasible; the next milestone is evidence from the experiments they help conduct.
