Quantinuum, in collaboration with JPMorganChase, Argonne National Laboratory, Oak Ridge National Laboratory, and the University of Texas at Austin, has announced a major milestone in quantum computing: the generation of certified quantum randomness using a commercial quantum computer. The results, published in Nature, demonstrate the first implementation of a certified-randomness-expansion protocol that produces more randomness than it consumes and cannot be achieved by classical methods. This protocol transforms Random Circuit Sampling (RCS)—originally proposed as a quantum supremacy benchmark—into a practical cryptographic tool, addressing a longstanding challenge in secure, verifiable entropy generation.
The experiment used Quantinuum’s H2 quantum processor, a 56-qubit trapped-ion system with high-fidelity gates and all-to-all qubit connectivity. Remote users submitted challenge quantum circuits to the device, which returned samples within a specified 2.5-second latency window. To confirm quantum origin, the same circuits were tested against classical simulation using DOE’s Frontier, the world’s most powerful supercomputer. The verification confirmed that at least 71,313 bits of entropy were mathematically certified as unpredictable, even under adversarial conditions, including hardware compromise or spoofing attempts.
Classical systems lack the intrinsic randomness of quantum mechanics and typically rely on physical entropy sources with limited scalability or verifiability. In contrast, quantum systems naturally exhibit nondeterministic behavior. Quantinuum’s implementation uses cross-entropy benchmarking, correlating the returned quantum samples with the expected circuit outputs. If the benchmark score exceeds a certain threshold and is returned within the latency limit, classical spoofing becomes infeasible—establishing the quantum-certified nature of the randomness.
This demonstration is a little different from another offering from Quantinuum called Quantum Origin. That approach uses a quantum computer to generate a quantum cryptographic seed and then uses a Bell test to verify the quality of the seed what was produced. This work builds on the protocol proposed in 2018 by Professor. Scott Aaronson and Shih-Han Hung, now demonstrated experimentally for the first time. The Aaronson/Hung protocol is based upon two steps. First, a users generates a challenge random circuit which they send to a remote quantum computer. Then the quantum computer returns very quickly a random number based about the random circuit it has received. Based upon the quickness of the response and the randomness quality of the returned number, a user can certify that the response was provided by a true quantum computer rather than a high performance classical computer.
The verified entropy can support applications in cryptography, differential privacy, secure multiparty computation, and simulation. Notably, the system achieves randomness certification without requiring trust in the quantum hardware—a critical feature for secure, cloud-based deployment. The collaboration leverages the combined computational power of multiple ExaFLOP-scale supercomputers, demonstrating an integrated quantum-classical workflow.
Quantinuum will commercialize the certified randomness protocol later in 2025 as part of its growing quantum cybersecurity suite, including Quantum Origin and the upcoming Helios platform, which will support at least 50 logical qubits. The results demonstrate a real-world application for quantum computing beyond classical capability, affirming the transition from theoretical utility to practical value in areas where verifiable entropy is mission-critical. The collaboration also reinforces JPMorganChase’s longstanding commitment to advancing quantum technologies for finance and beyond.
Read the official announcements and research from Quantinuum here, JPMorganChase here and here, and the Nature paper here.
March 27, 2025
Leave A Comment