IonQ generated buzz last year after CEO Niccolo de Masi expressed a bold ambition to Barron's: turning IonQ into the Nvidia of quantum computing. Now, that association is being taken a step further.

IonQ announced Wednesday that its latest quantum computer, Superion 256, will be the first system in Nvidia's quantum research center. Instead of operating as an isolated machine in a lab, the Superion 256 will act like a specialized co-processor alongside an Nvidia GPU cluster.

Using Nvidia's CUDA-Q software, quantum-specific tasks will be routed to the Superion 256 while Nvidia's GB200 GPUs handle heavy classical processing and artificial intelligence concurrently. It's a textbook example of hybrid computing, where quantum hardware is paired directly with existing supercomputing infrastructure.

CEO Niccolo de Masi described the announcement as "an exciting moment for compute and the history of compute." The development comes just weeks after IonQ unveiled the Superion range at its investor day, featuring chips built at its in-house foundry.

The Superion 256, which is slated to start shipping next year, is designed as an upgradable system that will progressively grow larger, featuring more quantum bits with time. According to de Masi, the chips are well suited to operate alongside Nvidia hardware in data centers, owing to their resilience against the mechanical vibrations and electromagnetic noise that typically disrupt quantum systems.

"We'll look back on this as a trailblazing moment when GPUs and QPUs (quantum processing units) can be integrated-not just today, but with a path to how we can do this with fault-tolerant machines that become exponentially more powerful," de Masi said.

Nvidia and IonQ have already partnered on research with Oak Ridge National Laboratory and the University of Tennessee to explore a combination of quantum algorithms and generative AI.

Timothy Costa, Nvidia's vice president and general manager for industrial and computational engineering, noted that the industry was "already trying to march forward" toward a hybrid model.

"It's going to look like an increase in heterogeneity and complexity of the systems that we already have today," Costa explained. "It might look, from a high level, very similar to what we have today."

Costa conceded that system-level integration will present major hurdles, akin to early efforts bringing GPUs into AI factories, adding that the desire to understand these pitfalls is what prompted Nvidia to build its quantum research facility.

There, researchers will address engineering challenges: determining which work belongs on each processor, how frequently they must exchange data, and the implications this has on system performance. These trials also reveal the financial costs, a crucial consideration for enterprises that will only adopt quantum systems if they deliver a clear speedup over existing infrastructure without blowing up budgets.

Integrating classical and quantum hardware is essential as the industry pushes toward fault-tolerant computing. Although the data being protected is quantum, the actual process of detecting and correcting errors relies on classical processing.

The market is already taking notice. IonQ shares surged more than 8% after hours on Tuesday after the company released a technical paper detailing a development in error correction. Error correction typically overwhelms classical processors, but researchers demonstrated how new architecture could handle errors in the background on a standard CPU, virtually eliminating system delays.

Nvidia aims to turn research like this into useful architectures and software for the broader quantum ecosystem, Costa said. He sees the partnership with IonQ as a major step in that direction.

The momentum extends beyond Nvidia and IonQ. Also on Tuesday, Microsoft unveiled a quantum research center of its own near the University of Maryland. Microsoft noted that the facility features dedicated workforce training spaces and brings its systems closer to federal partners for independent evaluation.