IBM, AMD and Oak Ridge Unveil Quantum Error Correction Platform

IBM, the U.S. Department of Energy’s Quantum Science Center at Oak Ridge National Laboratory (ORNL), and Advanced Micro Devices unveiled a co-design platform for quantum error correction on November 5.

The infrastructure co-locates IBM’s Nighthawk r2 quantum processors with ORNL’s Umbra high-performance computing system in Poughkeepsie, New York. The move aims to reduce the latency inherent in cloud-based hybrid workflows that typically connect quantum and classical computers, according to IBM.

Quantum error correction is a central challenge for practical, fault-tolerant quantum computing. Qubits are fragile and prone to decoherence from environmental noise, leading to gate errors ranging from one in 100 to one in 1,000 operations The Next Platform1. Logical qubits—groups of physical qubits protected by error-correcting codes—are essential for reducing these errors, but scaling them requires significant classical compute resources to run decoders in real time.

Classical decoders on AMD GPUs

The platform pairs IBM’s 120-qubit Nighthawk r2 processor with ORNL’s Umbra HPC system. Through IBM’s Quantum System API, researchers can access deep pulse-level capabilities, including mid-calculation measurement and real-time corrective signals newsroom.ibm.com5.

AMD’s hardware forms the classical foundation. The Umbra system uses AMD Instinct MI355X GPUs and EPYC 9005 datacenter CPUs to handle decoding optimization and software orchestration. Researchers will use these accelerators to train and evaluate decoders at scale, testing how GPU clusters facilitate quantum-classical workflows newsroom.ibm.com5.

IBM officials emphasized Nighthawk’s square lattice connectivity, a topology where most qubits connect directly to four nearest neighbors. This represents approximately a 60 percent increase in connectivity compared to IBM’s previous "heavy-hex" architecture and offers compatibility with standard surface codes for error correction, according to reporting by The Next Platform which cited IBM’s emphasis on these features The Next Platform1.

Travis Humble, director of the Quantum Science Center at ORNL, said the collaboration leverages ORNL’s high-performance computing strength to test new ideas and build the foundation for more reliable quantum computers. Jay Gambetta, Director of IBM Research, stated the question is now how to accelerate error correction to build systems scientists can use. Mark Papermaster, AMD CTO, described the platform as an important step toward scalable computing infrastructure.

Measuring against DOE's 2028 goal

The initiative supports the U.S. Department of Energy’s Quantum Genesis Mission. IBM says the DOE Quantum Genesis Q Competition targets 100 logical qubits running 100,000 hard operations by late 2028 newsroom.ibm.com5. The DOE website describes a goal of logical qubits numbering in the low hundreds in 2028, while its Request for Applications calls for systems capable of performing hundreds of millions fault-tolerant operations www.energy.gov3.

Current progress falls short of IBM’s framing of the target. In July, IBM and the University of Chicago demonstrated experiments involving up to 70 logical qubits and 438 hard operations, figures cited in the Nov. 5 IBM release newsroom.ibm.com5. IBM estimates closing this gap requires a roughly 230-fold increase in hard operation count, though Seetharami Seelam, IBM’s chief architect for classical system co-design, told The Next Platform the required increase is 238-fold The Next Platform1.

The co-design platform enables research into hierarchical error correction, a method layering two codes for protection. IBM research suggests this approach could reach a trillion logical operations at a physical qubit error rate of one in 1,000, provided sufficient classical compute resources manage the decoding layers The Next Platform1.

Work conducted on the platform may inform the development of IBM Quantum Starling, IBM’s first fault-tolerant quantum computer, which the company expects to deliver by 2029, according to an IBM blog post www.ibm.com6. The platform also connects with the Oak Ridge Leadership Computing Facility (OLCF) in Tennessee, creating a pathway to explore quantum computing as part of a broader AI and scientific computing infrastructure.

The announcement does not include independent test results validating the performance gains of the specific Nighthawk-Umbra integration. It also does not specify pricing or public access restrictions for external researchers wishing to utilize the Poughkeepsie facility. An ORNL announcement was unavailable for retrieval. The IBM blog states the platform is "open" to more researchers but provides no details on application procedures.