What if the power grid could predict exactly how much electricity your street needs tomorrow, hour by hour, instead of guessing from last year’s averages? That question just got a real answer. A Sydney-based quantum computing company teamed up with Schneider Electric, and their quantum energy forecasting system is already running up to 41 percent more accurate than the models utilities use today.
Silicon Quantum Computing, or SQC, builds a chip called Watermelon. It is not a lab experiment locked behind glass. It slots into a normal server rack, the same kind data centers already use. SQC just advanced to Stage 2 of Australia’s Critical Technologies Challenge Program, working with Schneider Electric and UNSW Sydney. The government backed this phase with another A$3.6 million.
A Chip That Finds Patterns Classical Computers Miss
Watermelon does not replace the computers utilities already run. It works beside them. The chip generates quantum features: extra signals buried in messy, fast-changing data. Classical software tends to miss these patterns. Blend them with ordinary data, and the forecast gets sharper.
SQC founder and CEO Michelle Simmons put it simply:
“Quantum processors would work alongside CPUs and GPUs to deliver real-world performance gains.”
Why Quantum Energy Forecasting Beats the Old Models
Stage 1 tested the system on real household data for a full year. The results were not a one-time fluke.
| Metric | Improvement Over Classical Forecasting |
|---|---|
| Average accuracy gain | 20% |
| Best-case accuracy gain | 41% |
Stage 2 scales that work up to hundreds of homes. It folds straight into Schneider Electric’s day-to-day operations.
Picture Your Power Bill, Minus the Guesswork
Picture a nurse in Brisbane who works night shifts. She gets home at 2 a.m., plugs in her electric car, and switches on the air conditioner before bed. Multiply her by thousands of households, each on its own schedule, and the grid gets genuinely hard to predict.
Three things drive that unpredictability:
- Rooftop solar that floods the grid with power at noon and vanishes by dinner
- Electric vehicles charging on their own unpredictable schedule, often overnight
- Home batteries that store and release power in ways old forecasting models never had to handle
Utilities already handed a quantum computer a similar job, untangling the grid in Chattanooga. The logic matches what SQC and Schneider are testing now: guess wrong about tomorrow’s demand, and utilities pay premium prices for emergency backup power. That expense flows back into your rate.
Shave 20 to 41 percent off that error, and even a conservative slice, say 1 to 3 percent off a typical power bill, adds up over a year. It also means fewer strained afternoons when the grid scrambles during a heat wave.
SQC already ran a similar pilot with Telstra on network demand. The Economist noticed Watermelon too, pointing out that it cuts energy use instead of adding to it. That matters anywhere AI’s power appetite is already a headache, not just in Australia.
Britain tackled a similar forecasting problem from another angle. A free solar-battery pilot there is already slashing low-income household bills by matching supply to demand more precisely. Pair programs like that with sharper quantum forecasting, and the savings on both ends start to compound.
Watch what happens when Stage 2 results land. If quantum forecasting holds up across hundreds of real homes, the next pilot like this probably will not be in Sydney. It might be your own utility. Ask them how they forecast demand on your street. Today the honest answer is “we guess, pretty well.” Give it a few years, and it might be “we used quantum.”