Second edition Watt Matters in AI Coming in November 2026!
Research: quantum computing could reduce AI’s growing energy demand
AI is driving up demand for computing power and electricity. But quantum computing could help make future AI systems more energy efficient. That is the conclusion of new research by quantum computing company IonQ, together with researchers from QuantumBasel and the Center for Quantum Computing and Quantum Coherence.
The study looks at one of the biggest challenges facing AI: its growing energy consumption. Training and running large AI models requires huge amounts of computing power. According to the researchers, traditional hardware such as GPUs and CPUs may struggle to keep up with future AI workloads without significantly increasing energy use.
The researchers investigated whether quantum computers can take over parts of the AI process in a more efficient way. Instead of replacing traditional computers completely, they propose a hybrid system. In this approach, classical computers perform standard AI tasks, while a quantum processor handles complex calculations where it can offer an advantage.
The team tested their approach using IonQ’s 36-qubit trapped-ion quantum computer. They measured the energy use of real quantum hardware and compared it with classical computer simulations. The results showed that energy consumption increased much more slowly on the quantum system. According to the researchers, quantum computers could become more energy efficient than classical simulations at around 34 qubits, although the exact point depends on the hardware and the application.
The team also developed methods to reduce errors caused by noise in current quantum systems. With these techniques, the researchers achieved a 24 percent reduction in errors compared with purely classical AI models in the tested tasks.
The study suggests that quantum computing could become a useful addition to AI infrastructure sooner than expected. Instead of waiting for fully developed quantum computers, current systems may already help reduce the energy costs of advanced AI applications.