Newsroom
17 November, 2026 / News / AI / Tags: quantum, decoding, ibm, codesign, nvidia

New decoding algorithm delivers up to 13,000 times faster performance while partners build modular systems to advance fault-tolerant quantum computing research
Nvidia, IBM and the Fermi National Accelerator Laboratory announced on November 16 a set of tools and a collaborative platform designed to accelerate progress in quantum error correction. The efforts center on a new decoding algorithm and a modular codesign environment that pairs classical accelerated computing with quantum processors.
Nvidia introduced TCBP-JOBS, a decoding algorithm released as part of its CUDA-Q QEC 0.9.0 library. The company reported that the algorithm achieves decoding speeds up to 13,000 times faster than existing leading methods. In research applications, this reduces the time required to explore new quantum computing designs from 12 weeks to approximately 45 minutes.
The algorithm combines trajectory-calibrated belief propagation with GPU-parallel jittered ordered statistics decoding beam search. It targets the classical computation required to identify and correct errors in quantum processors. Nvidia stated that the advance improves the efficiency of physical qubit use and supports the development of fault-tolerant systems.
Krysta Svore, Nvidia’s vice president of applied research for quantum computing, noted that realizing useful quantum computers for scientific work will require progress across multiple areas.
Alongside the algorithm release, Nvidia, IBM and Fermilab outlined plans for a modular quantum computing codesign platform. The setup will place Nvidia DGX B200 systems in close proximity to IBM quantum computers, including systems based on the IBM Quantum Nighthawk processor.
Researchers plan to use the environment to develop and evaluate error-correction codes, decoding methods and quantum-classical applications. Initial focus areas include high-energy and nuclear physics. Fermilab will contribute expertise in scientific applications and later integrate its ultracoherent superconducting cavity hardware for studies of quantum memory, entanglement buffering and bosonic error correction.
IBM will supply superconducting quantum processors and related software. Nvidia will provide accelerated computing resources for decoding, simulation and real-time feedback. The partners described the platform as a step toward establishing technology, software and user communities needed for future scalable quantum facilities.
IBM describes its Quantum Nighthawk as its fastest quantum processor, operating with 120 qubits. Quantum systems remain constrained by error rates that limit the complexity and duration of reliable computations. Error correction remains a central technical barrier to broader practical use.
Discussions of long-term cryptographic implications continue in parallel. A 2022 University of Sussex study estimated that breaking Bitcoin encryption within one day would require roughly 13 million physical qubits. Early Bitcoin addresses, including those associated with Satoshi Nakamoto’s early mining activity, expose public keys on the blockchain and would face elevated risk if large-scale fault-tolerant quantum computers emerge.
Developers have advanced protective proposals. In 2024, Bitcoin developer Hunter Beast introduced BIP-360, which defines a new output type intended to protect coins from quantum attacks. In July 2025, Casa co-founder Jameson Lopp outlined steps to migrate holdings from legacy signatures to quantum-resistant address formats. In November 2024, the U.S. National Institute of Standards and Technology recommended discontinuing vulnerable public-key algorithms, including elliptic-curve cryptography, by 2035.
At the time of the announcements, Bitcoin traded near $64,580.
The new tools address the classical side of error correction, where decoders must process measurement data quickly enough to keep pace with quantum operations. Faster research cycles allow more thorough exploration of code designs and noise models. The codesign platform is intended to tighten the feedback loop between quantum hardware, classical acceleration and scientific application development.
Partners indicated that the work complements other recent collaborations focused on quantum-classical integration. The combined announcements mark a coordinated push to move quantum systems closer to reliable, large-scale operation for both scientific and potential commercial workloads.







