Oak Ridge Starts Weaving Together A Quantum, Classical HPC, And AI System Stack
Jump to main content
Search
NEXTPLATFORM AD
HPC
Oak Ridge Starts Weaving Together A Quantum, Classical HPC, And AI System Stack
Jeff Burt
Jeff<br>Burt
Published<br>thu 21 May 2026 // 14:18 UTC
It is common understanding that as quantum computing gets<br>its feet under itself, it will work hand-in-hand with classical supercomputers<br>and leverage what rapidly evolving AI tools can offer to begin solving some<br>thorny calculations that the largest HPC systems may be unable to address.<br>We’ve written about this direction quantum is heading in and steps major<br>players already are taking to address what IBM calls “quantum-centric<br>supercomputing.”<br>Big Blue earlier this month demonstrated<br>the largest simulation of molecules performed with quantum hardware,<br>pairing its 156-qubit Heron r2 processors running in IBM quantum systems at the<br>Cleveland Clinic and at RIKEN in Japan in tandem with two classical<br>supercomputers, the Fugaku and Myaybi-G systems. In March, IBM rolled out<br>a reference<br>architecture for integrating quantum and HPC systems.
NEXTPLATFORM AD
Nvidia is developing technologies to more<br>tightly link supercomputers with quantum systems, and the need for such a<br>pairing is being<br>pushed at a national level.<br>The challenge now is finding ways to make these systems work<br>together as seamlessly as possible, which means not only linking the hardware<br>but also addressing everything from algorithms to software to the role AI plays<br>in the mix. What jobs – or what portion of jobs – will run on classical<br>supercomputers rather than quantum systems, and what mechanisms will determine<br>how they move back and forth.<br>That is among the priorities of the quantum computer work<br>being done at the US Department of Energy’s Oak Ridge National Laboratory,<br>according to Tom Beck, section head for Science Engagement for the National<br>Center for Computational Sciences (NCCS) at the national research facility in<br>Tennessee.
Tom Beck, left, stands with fellow ORNL researchers Sarp Oral and Rafael Ferriera da Silva in front of the lab's Frontier supercomputer.
Beck, who also is the section head or Oak Ridge’s quantum-HPC<br>unit, tells The Next Platform that a key area Oak Ridge scientists are<br>looking into is the ongoing convergence of supercomputers, AI, and quantum,<br>what he calls the dominant areas the next era of HPC. Oak Ridge is home to Frontier, first exascale-class system in the United States. Comprised of HPE’s Cray<br>EX235A systems powered by AMD’s custom 64-core Epyc 2GHz processors and Instinct<br>MI250X GPUs and linked through the hardware maker’s Slingshot-11<br>interconnect, it still ranks four years after rollout as the second-fastest<br>system on the Top500 list.
NEXTPLATFORM AD
As for AI, it’s “exploding in importance across business and<br>science,” Beck says, noting the DOE’s Genesis<br>Mission initiative to build an AI-driven, integrated compute platform to<br>accelerate scientific discovery in energy, national security, and technology. The<br>program, which in March received<br>$293 million that interdisciplinary teams can vie for to tackle some of the<br>core<br>26 challenges outlined by the DOE, connects all 17 national labs with private<br>sector companies in AI and supercomputing, like Microsoft, Nvidia, and OpenAI.<br>“Quantum computing is at an earlier stage, but it's<br>developing rapidly and we are trying to figure out how to link quantum<br>computing to HPC, and quantum computing at this stage can be viewed as an<br>accelerator similar to GPUs 25 years ago,” Beck says. “Quantum computing allows<br>you in principle to solve some exponentially scaling problems in a polynomial<br>amount of time. In other words, you can solve problems that you could never<br>access even on a machine like Frontier. Those problems are not that many at<br>this time. There could be encryption and national security-type problems.<br>That's definitely a big application.”<br>Oak Ridge scientists have been working on the details of a<br>hybrid HPC-quantum environment for several years. It not only houses Frontier<br>but also the Quantum Computing User Program, which opens time on privately owned<br>quantum processors to support quantum studies, and it leads the DOE’s Quantum<br>Science Center.<br>In a study in 2024, Beck and other ORNL scientists proposed<br>such ideas<br>as creating quantum test beds to work with a range of technologies and pair<br>those test beds with classical machines. They also recommended a high-speed<br>network be developed to connect classical HPC systems to their quantum<br>counterparts.
NEXTPLATFORM AD
Getting Quantum And Classical to Work Together<br>Such technologies would be useful as Oak Ridge scientists<br>continue to explore how the two types of systems can work together. As an<br>example, Beck points to a software stack on a supercomputer may be linked to a<br>smaller set of GPUs that control the quantum device and provide access to it so<br>that some parts of the problem are offloaded onto the HPC system.<br>“We do the quantum sampling,...