NVIDIA, IBM, and Fermi National Accelerator Laboratory announced a joint quantum computing co-design platform that will pair NVIDIA DGX B200 systems with IBM quantum computers, including those powered by IBM's Nighthawk processor, to support error-correction and applications research in high-energy and nuclear physics. IBM describes Nighthawk as its most powerful quantum processor, with 120 qubits and speeds of up to 100,000 samples per second. The collaboration involves the United States Department of Energy's Fermi National Accelerator Laboratory.
Separately, NVIDIA extended the Slurm workload manager to natively schedule quantum processing units alongside classical GPUs. The QPU support uses a single vendor-defined interface file reusable across supercomputing center deployments, removing the need for custom per-site integration work. Alice & Bob, IBM, IQM, Pasqal, Quantinuum, and MQSC have adopted the extension, which supports the QRMI and QDMI interfaces. Slurm 26.11, carrying this support, is slated to be available later in November 2026.
On the software side, NVIDIA's CUDA-Q platform includes a new TCBP-JOBS decoder for quantum error correction. NVIDIA claims the decoder cuts error-correction research computation time from 12 weeks to 45 minutes, and states a goal of reducing quantum decoding latency from roughly 300 microseconds to 10 microseconds. A simulation on Jülich's JUPITER supercomputer demonstrated that magic state cultivation can reduce the operations needed for one logical quantum step by tenfold. Fermilab used CUDA-Q to complete studies in three weeks that would otherwise have taken five months, and QunaSys ran simulations of more than 300 million quantum operations on 1,024 NVIDIA Hopper GPUs using cuQuantum, reaching chemical accuracy for a four-hydrogen-atom model.