The Challenge
DARPA posed the question: Does any meaningful, demonstrable quantum advantage exist—and if so, for which problems, on what hardware, and under what resource constraints? To answer this, DARPA launched the Quantum Benchmarking (QB) program, an ambitious effort aimed at establishing the first rigorous, end-to-end methods for quantifying where quantum computers could deliver practical, beyond-classical performance. Achieving this requires comparing a rapidly evolving landscape of quantum devices, estimating the classical and quantum resources required to solve targeted problems, and understanding how soon any genuine quantum advantage could be realized.
Demonstrating a transformative quantum advantage is extraordinarily difficult. Quantum systems are noisy, fragile, and rapidly evolving, while classical algorithms and hardware continue to improve at a remarkable pace. DARPA’s challenge is to determine—not hypothetically, but with rigorous evidence—whether any quantum algorithm on any foreseeable hardware can truly outperform the best classical methods on a problem that matters. This requires navigating massive uncertainties in hardware capability, scaling behavior, noise modeling, and resource requirements, all while comparing fundamentally different computational paradigms on equal terms.
Our Approach
Zapata was selected by DARPA for all Technical Areas of the Quantum Benchmarking (QB) program, underscoring the breadth of our capabilities in quantum algorithms, resource estimation, and benchmarking methodology. Together with partners across government, industry, and academia, we examined domains often cited as leading candidates for quantum impact: quantum chemistry simulation, advanced materials discovery, combinatorial optimization, and machine-learning–driven inference tasks. Across these areas, Zapata provided the integrated algorithmic, modeling, and benchmarking work needed to evaluate whether any pathway to transformative quantum advantage exists and to give DARPA a grounded, evidence-based framework for making that determination.
Technical Area Contributions
TA1 – Problem & Algorithm Definition:
Zapata helped identify and define the high-value application problems most likely to yield transformative quantum advantage. We also explored the quantum algorithmic pathways that could plausibly tackle these challenges, evaluating the state of the art, identifying promising algorithm families, and mapping how they might scale with future hardware. The goal was to ensure the program focused on real, consequential workloads, not toy problems, and established a solid foundation for the more detailed benchmarking and resource estimation work that followed.
TA1.5 – Benchmark Construction & Validation:
TA1.5 was an intermediate technical area added after the start of the Quantum Benchmarking (QB) program to fill a crucial gap between high-level problem definition (TA1) and full resource estimation (TA2): identifying promising quantum use cases and actually measuring their performance. Zapata transformed the high-value problems and potential algorithmic pathways, into concrete, well-specified benchmark tasks with clear performance targets. This made it possible to compare classical and quantum approaches on equal footing and drive the program toward actionable results.
TA2 – Resource Estimation & Performance Modeling:
Zapata quantified the computational resources required to solve each benchmark problem. This included noise modeling, scaling analyses, implementation estimates, and comparisons against state-of-the-art classical algorithms and hardware. We led the development of BenchQ, a comprehensive toolkit for estimating the hardware resources required for fault-tolerant quantum computation, providing components such as a graph-state compiler, distillation factory and decoder models, an ion-trap architecture model, and implementations of selected quantum algorithms.
