Insider Brief
- QCentroid has announced QCentroid Missions, a new QuantumOps capability that helps enterprise teams evaluate computational approaches against business objectives and defined validation criteria.
- The system uses agents to test classical, quantum-inspired, GPU-simulated and quantum processing unit (QPU) approaches, with recommendations based on benchmarking evidence and the option to retain an existing classical solution.
- QCentroid Missions uses NVIDIA CUDA-Q for accelerated circuit simulation and is being demonstrated at NVIDIA GTC Berlin from 20–22 October 2026, with rollout to QCentroid Platform customers planned from Q4 2026.
PRESS RELEASE — QCentroid today announced QCentroid Missions, a new goal-driven capability in its QuantumOps platform designed to take enterprise teams from a measurable business objective to validated evidence for deciding whether, and how, to evolve an existing computational solution. QCentroid Missions is being demonstrated this week at NVIDIA GTC Berlin.
Rather than starting with a quantum algorithm or a predefined technology, a Mission starts with an outcome such as reducing execution time, improving solution quality or reaching a target business metric. Teams define the objective, the computational and operational limits, and the validation gates that must be cleared. Expert agents in QuantumOps then run the experimentation cycle, evaluating potential evolutions of the current solver across classical, quantum-inspired, GPU-simulated and QPU-ready approaches.
The resulting recommendation may be classical, hybrid or quantum, including a recommendation to retain the current classical solution when the evidence supports it.
“Quantum teams are small, while the number of questions coming from the business continues to grow,” said Carlos Kuchkovsky, CEO and co-founder of QCentroid. “With Missions, teams no longer have to manually drive every stage of an experiment. They define what they want to improve, how far the system is allowed to go and what counts as proof. The agents run the cycle, and the team keeps the decision.”
“Useful workloads in future quantum supercomputers will draw on CPUs, GPUs and QPUs” said Sam Stanwyck, director of quantum product at NVIDIA. “QCentroid Missions builds on CUDA-Q to help enterprise teams prepare for this paradigm.”
From Baseline to Production Evidence
A Mission begins with the organisation’s current solver and evaluates potential evolutions against it using a common benchmarking contract. The objective is not to prove that a quantum approach should be used, but to determine which computational approach best meets the business goal under the agreed constraints.
Take collateral allocation as an illustrative example. A Mission could compare an exact optimisation approach, quantum-inspired annealing and a QAOA-based solver under the same evaluation framework. As the problem grows, a time-limited exact solver may struggle to return a feasible solution, while alternative approaches may still produce feasible allocations of comparable quality. The Mission iterates on each implementation before returning the benchmark evidence. That lets the team judge which solver path warrants further investment, rather than assuming in advance that a quantum approach should prevail.
Validation is built into the workflow. In a risk-modelling scenario, for instance, if an initial performance result appears stronger than expected, the system’s validation layer is designed to challenge it, independently recompute it and, where necessary, revise it before it is reported. Only results that pass the validation gates progress to a decision.
Goals, Limits and Human Validation
Each Mission combines a measurable goal, limits covering areas such as compute budget, GPU hours, time, code, data and permitted execution backends, and validation gates defining the checks and human approvals required before results can progress or be published.
Within those boundaries, QuantumOps agents perform discovery, feasibility analysis, solver design, implementation, benchmarking, debugging and iteration. People remain responsible for defining what is being attempted, which resources the Mission may consume and what level of evidence is required for a decision.
Built for Hybrid Computing
Circuit simulation in QCentroid Missions runs by default on NVIDIA accelerated computing through the QCentroid Launchpad integration with NVIDIA CUDA-Q.
QuantumOps uses a common experiment lifecycle across classical, quantum-inspired and quantum approaches, while QPU execution remains a gated and QPU-agnostic option when the evidence justifies its use. This allows compute resources to be selected as part of the application engineering process rather than forcing teams into separate experimentation environments for different solver families.
Availability
QCentroid Missions is being demonstrated at NVIDIA GTC Berlin, 20–22 October 2026, and will roll out to QCentroid Platform customers from Q4 2026.