Silicon Quantum Computing (SQC), a Sydney company, and Schneider Electric say a quantum-enhanced method forecast next-day home energy demand better than a classical benchmark in a 12-month test. The numbers are the companies’ own, and no outside group has checked them.
At a glance
| Item | Detail |
|---|---|
| Reported gain | 20 percent on average, up to 41 percent at best, over a classical benchmark (company-reported) |
| Product | Watermelon, a quantum feature generator for machine learning |
| Availability | SQC says by cloud access or purchase |
| Status | Stage 1 done; Stage 2 planned, with no timeline |
| Funding | A$3.6 million (about US$2.5 million) under Australia’s Critical Technologies Challenge Program, with UNSW Sydney |
How Watermelon fits in
Watermelon doesn’t replace the computers utilities already run. It works beside them. SQC says it generates quantum features, extra signals that feed classical forecasting models, and the companies say blending these features with ordinary data improved the forecast. The announcement doesn’t describe the method, the benchmark model or the error metric. SQC founder and CEO Michelle Simmons put it simply:
“Quantum processors would work alongside CPUs and GPUs to deliver real-world performance gains.”
In Stage 1, the team applied Watermelon to next-day forecasting data over 12 months, though Quantum Computing Report describes the forecast as simulated. SQC and Schneider Electric then advanced to Stage 2 of the Critical Technologies Challenge Program. They say they’ll expand modelling to hundreds of homes across Australia and build Watermelon into Schneider Electric’s AI workflows.
Home energy is hard to forecast. Rooftop solar floods the grid at noon and vanishes by dinner, electric vehicles charge on their own schedules, often overnight, and home batteries store and release power in ways old models never had to handle. (As an illustration, not a real person: a nurse who gets home at 2 a.m., plugs in her car and switches on the air conditioner, multiplied by thousands of households.) We covered a related grid effort, a quantum project in Chattanooga.
SQC ran a similar 12-month trial with Telstra on network metrics, announced October 13, 2025. The companies reported accuracy on par with a deep learning model, with no figures published. A Quantum Zeitgeist repost of SQC’s claims says Watermelon fits a standard 19-inch rack and could reduce AI’s environmental impact. It gives no energy figures, so treat that as a company claim.
Who Benefits and How
- Utilities come first. When one misjudges demand, it may need to buy extra power at short notice, so better forecasts could cut those costs. Schneider Electric’s Colette Munro says the partnership aims to help manage this complexity and lower consumer costs.
- Households could benefit only indirectly. The companies have published no savings figure, and a smaller forecast error doesn’t automatically mean a smaller bill.
- Nothing is available to customers today. Stage 2 will test the method on hundreds of homes in Australia, and Watermelon itself can be bought or accessed by cloud by those who build forecasting models.
- Other programs try to cut bills from a different angle, such as a solar-battery pilot in Britain.
What to keep in mind
The 20 and 41 percent figures come from the companies’ press release, and The Quantum Insider’s report contains no independent verification. The announcement doesn’t define “accuracy,” name the error metric or describe the classical benchmark. The 20 percent is an average and the 41 percent a maximum, with no spread reported, so most days could sit well below the peak. Quantum Computing Report describes the Stage 1 forecast as simulated and says it is unclear whether results came from the physical device or from emulation. Stage 1 household numbers aren’t given, and Stage 2 is described only as “hundreds of homes.” If the gains hold across those homes and outside experts can check them, other utilities may take interest. Until then, treat the figures as company-reported.
Quick answers
What did SQC and Schneider Electric announce?
They advanced to Stage 2 of Australia’s Critical Technologies Challenge Program and received A$3.6 million. They report that Stage 1 showed an average 20 percent forecasting gain, with gains up to 41 percent.
What is Watermelon?
SQC calls it a quantum feature generator that works with classical machine learning. SQC’s product page says it is available to purchase and through cloud access. The page cites no benchmarks.
Will this lower my power bill?
Not yet. The companies have published no savings figures.
Has SQC done something like this before?
Yes. In an exploratory 12-month trial announced October 13, 2025, SQC and Telstra reported Watermelon matched a deep learning model on network forecasting. That release gave no numbers.
Sources
- The Quantum Insider: SQC and Schneider Electric Advance Quantum-Enhanced Energy Forecasting (Oct 02, 2026)
- Quantum Computing Report: SQC and Schneider Electric Advance Energy Grid Forecasting (Oct 03, 2026)
- Silicon Quantum Computing: Products
- The Quantum Insider: Telstra and SQC Explore Smarter Network Prediction (Oct 13, 2025)
- Quantum Zeitgeist: Watermelon AI’s Energy Demands Shift Quantum (Jul 29, 2026)
