IonQ and Nvidia report AI method for faster quantum-circuit design

IonQ ($IONQ), Oak Ridge National Laboratory, Nvidia ($NVDA) and the University of Tennessee reported research using a trained AI model to generate quantum optimization circuits directly. In the benchmark, circuit-finding time stayed near 28 seconds as the tested subproblems grew, while the comparison method increased from roughly 34 seconds to more than 11 minutes.

The model was trained on high-performing circuits and generated candidates that were simulated, scored and selected for the solution. The reported result concerns reducing the repeated tuning needed to design circuits. It compares two quantum-circuit generation approaches, not quantum computers against the best classical solvers, and does not establish commercial quantum advantage. The work is a potentially useful development milestone, rather than a customer contract or revenue announcement.

Sources:

• IonQ — joint research announcement