IonQ intends to place the Superion 256 system in the NVIDIA Accelerated Quantum Research Center, where it will become the company's first quantum processor installed in NVIDIA's infrastructure. The 256-qubit QPU is planned to be directly linked to the GB200 NVL72 computing system via the NVIDIA NVQLink interconnect, and the open CUDA-Q platform will be used to manage task distribution. This configuration is expected to allow exploration of the interaction between quantum computing and GPU-accelerated infrastructure and AI models. The aim is not to replace classical supercomputers with quantum systems, but to jointly perform various stages of computational tasks on specialized components.
As part of the project, IonQ and NVIDIA intend to develop software for hybrid computing, test scalable system architectures, and publish research results. Potential scenarios include optimizing investment portfolios and assessing financial risks, modeling materials, and computational chemistry, including tasks related to drug compound development. Special attention will be paid to how computations are distributed between the QPU and classical accelerators. IonQ also links the project to its long-term architecture aimed at increasing the number of qubits and creating fault-tolerant quantum systems. The company points to vertical integration of production, including the use of SkyWater CMOS technologies, as one of the elements of this strategy.
The Superion platform was introduced by IonQ on September 8, 2026. The system is already available for order, with the first deliveries to clients scheduled for 2027. The placement of Superion 256 in the NVIDIA research center is also planned for the same period. For the practical value of such a connection, the key question will be not only the number of qubits but also the efficiency of data exchange between the quantum processor and classical computing resources. The hybrid architecture assumes that different parts of a single task will be executed on the most suitable equipment, and the software layer must coordinate this process. Therefore, future experiments will allow assessing which scientific and applied scenarios truly benefit from the integration of QPU, GPU, and AI tools in a single computing environment.
Related articles
