AIBRIEF
After a year long trial cut next day energy forecast errors by an average of 20%, Silicon Quantum Computing and Schneider Electric have advanced to Stage 2 of an Australian government program, with A$3.6 million in fresh funding.
A Sydney-built quantum chip is now sitting inside Schneider Electric's forecasting pipeline, where a 12-month trial cut next-day energy forecast errors by an average of 20% against a classical baseline, with gains up to 41%. The result, announced via PR Newswire on 2 October 2026, lands as grids struggle to absorb more rooftop solar, home batteries, and EVs whose output households can flex on their own.
SQC's Watermelon chip is a feature generator, not a replacement for classical AI. It produces quantum-derived signals that feed the same machine learning workflow Schneider Electric already runs, sitting alongside CPUs and GPUs. SQC describes Watermelon as a quantum feature generator for classical models, adding a layer of signals the classical side tends to miss on its own.
Stage 2 brings A$3.6 million (US$2.5 million) in fresh funding through the program and expands the work to hundreds of Australian homes, with the University of New South Wales (UNSW) Sydney continuing as research partner. The chip will be integrated directly into Schneider Electric's AI workflows. It is the first time a quantum feature generator ships inside a global energy vendor's production-adjacent stack, according to The Quantum Insider's coverage of the announcement.
The 20% and 41% figures are company-reported, and the classical benchmark definition has not been published. Stage 2 will show whether the gain generalizes.