IonQ, ORNL, Nvidia, and University of Tennessee Develop DQAOA-GPT for Quantum Optimization

A groundbreaking collaboration among IonQ, Oak Ridge National Laboratory (ORNL), Nvidia, and the University of Tennessee has resulted in the development of DQAOA-GPT, a hybrid framework designed to improve quantum optimization processes. This innovative system combines quantum approximate optimization with generative AI to tackle complex combinatorial problems more effectively than traditional methods.

The core advancement of DQAOA-GPT lies in its ability to synthesize efficient quantum circuits in a single attempt, eliminating the need for extensive iterative loops that typically characterize quantum optimization. This approach significantly reduces the computational costs associated with evaluating quantum circuits, which is particularly crucial given the limitations of current noisy intermediate-scale quantum (NISQ) devices.

In tests involving dense Higher-order Unconstrained Binary Optimization (HUBO) instances, the DQAOA-GPT framework demonstrated substantial cost savings compared to conventional variational approaches, achieving high-quality solutions with fewer circuit evaluations and shallower circuits.

The research was recognized at IEEE Quantum Week (QCE 2026), where the paper detailing DQAOA-GPT earned third place out of 857 submissions. IonQ's strong performance at the conference included four Best Paper awards, highlighting its commitment to advancing hybrid quantum-classical architectures.

Each collaborator contributed unique strengths to the project: Nvidia provided accelerated computing infrastructure, IonQ offered its high-fidelity trapped-ion quantum hardware, and ORNL and the University of Tennessee brought expertise in high-performance computing and quantum information science.