Quantum Benchmarking Initiative (QBI)
The Quantum Benchmarking Initiative (QBI) of the Defense Advanced Research Projects Agency (DARPA) is an important program designed to thoroughly assess the development and promise of quantum computing technologies. Its main objective is to ascertain whether, by 2033, any quantum computing technique can accomplish “utility-scale” operation, in which its computational value surpasses its cost. This project is seen as an important step in transforming quantum computing from an experimental and theoretical area to a useful and industrially applicable technology.
How It Works (Stages)
Every effective performer endeavour will move through the three crucial stages of the QBI’s structure:
Stage A: Concept Description (6 months)
- Businesses offer comprehensive technical proposals to show that developing utility-scale quantum systems is feasible.
- The goal of this stage is to describe a utility-scale quantum computer concept that has a reasonable chance of becoming a reality in the near future.
- Each startup may receive up to $1 million in funding for Stage A.
- Businesses need to respond to questions on how they want to construct their quantum computer, why they think they can do it, how it will affect the world, and what their objectives are for the upcoming year. As a participant, IonQ will help define utility-scale performance by reviewing different problem sets and use cases that call for large-scale machines through DARPA.
Stage B: Research & Development Plan (12 months)
- Companies who make it through Stage A will go on to Stage B to further examine their plans for research and development and to specify how they will gauge their progress.
- At this point, businesses must outline a research and development strategy that can create a utility-scale quantum computer, including the risks involved, the methods they aim to take to reduce those risks, and the prototypes they will need.
Stage C: Verification & Validation (variable, nominally 36 months)
- Businesses that advance through Stage B will go on to Stage C, where they will construct their hardware and submit to a DARPA team’s stringent independent verification and validation.
- For these assessments, DARPA has put together a highly skilled test and evaluation team that will make use of a number of federal and state test facilities.
- Determining if a company’s technology is effective and confirming that the suggested system can be built as intended and run as intended for real-world deployment are the ultimate goals of Stage C.
You can also read Poor Man’s Majorana: Spin-Exchange Spillover Effect Revealed
Architecture of Quantum Benchmarking
The methodical process of evaluating the hardware, software, and algorithms of quantum computers is known as quantum benchmarking. To assess performance at various quantum stack levels, the architecture uses a multi-layered approach:
Component-level: This layer evaluates the performance of individual qubits and gates by looking at things like coherence times and fidelity (accuracy).
System-level: This assesses the quantum processor as a whole, paying particular attention to the way mistakes spread over several qubits.
Software-level: This layer evaluates the effectiveness of the quantum compiler and the different error-reduction strategies used.
Application-level: This is the highest level, where benchmarks are grounded in real-world issues from domains such as optimization, machine learning, and quantum chemistry, offering some degree of usefulness.
Types of Quantum Benchmarks
Quantum benchmarks are categorized based on what they measure and their complexity:
Low-Level Benchmarks (Diagnostics): The primary goal of low-level benchmarks (diagnostics) is to identify certain hardware problems. Quantum Process Tomography, which offers a thorough description of a particular gate or process, and Randomized Benchmarking (RB), which calculates the average error rate of quantum gates, are two examples.
Volumetric Benchmarks: These give the total capability of a quantum computer as a single number. Quantum Volume (QV), which quantifies the greatest “square” circuit (equal number of qubits and circuit depth) that a machine can operate with adequate fidelity, is a well-known example.
Application-Driven Benchmarks: The purpose of application-driven benchmarks is to evaluate a quantum computer’s performance on particular, realistically applicable tasks. They are essential for determining whether a machine is actually beneficial for practical uses. Benchmarks for training quantum machine learning models, resolving optimization issues, and modelling molecular behavior are a few examples.
Quantum Supremacy/Advantage Benchmarks: These benchmarks, like Random Circuit Sampling (RCS), are intended to show that a quantum computer can do a computational task noticeably quicker than the top-performing classical supercomputer.
You can also read Infleqtion to Go Public via with Churchill Capital Corp X
Features of QBI
The QBI highlights a number of important features:
Focus on Utility-Scale: By 2033, the main goal is to develop and validate “utility-scale” quantum computing, in which the benefits exceed the drawbacks.
Diverse Qubit Technologies: DARPA has chosen businesses that use a variety of qubit technologies, such as photonic, trapped ion, superconducting, and neutral atom qubits. Atom Computing (neutral atoms), Diraq (semiconductor spin qubits), IBM (superconducting qubits), IonQ (trapped ions), Quantinuum, QuEra (neutral-atom platform), and Xanadu (photonic qubits) are a few instances of publicly traded enterprises.
Not a Competition: QBI is intended to evaluate the commercial quantum landscape and assist all conceivable avenues towards revolutionary quantum computing, not to serve as a competition among performers.
Independent Verification and Validation (IV&V): One important aspect is that the performers’ routes to a utility-scale quantum computer be verified and validated by an impartial third party.
Government Communication: QBI will actively notify pertinent U.S. government entities interested in using or purchasing this technology of the positive outcomes of its verification and validation activities.
Partnerships: Building on their long-standing alliances to advance commercially scalable quantum computing, Quantinuum’s Stage A endeavor includes relationships with Microsoft and NVIDIA.
Advantages
The implementation of quantum benchmarking, particularly through initiatives like QBI, offers several advantages:
Standardization: Benchmarks offer a consistent method for evaluating various quantum computers and efficiently monitoring their development over time.
Transparency: By providing unbiased, verifiable performance data, they encourage transparency in the realm of quantum computing.
Guidance for Development: In order to attain utility-scale quantum computing, benchmarks assist researchers and developers in concentrating on the most important areas for advancement.
Reality Check: In order to preserve investor confidence, the QBI acts as a vital reality check for the quantum computing industry, working to dispel myths and uncertainties and control the prevailing hype.
Inspiration for Others: In order to create distinct development paths, the project may encourage other countries that are actively researching quantum technology, like India, to start comparable quantum benchmarking projects.
Disadvantages
Despite its benefits, quantum benchmarking faces certain disadvantages:
Manipulation: Benchmarks may be “gamed” or tailored to produce particular outcomes, which could result in inaccurate comparisons between various quantum systems.
Limited Scope: The capabilities of a quantum computer may not be fully represented by a single benchmark; a machine that performs well on one kind of benchmark may not do well on another.
Complexity: It is intrinsically difficult to create strong and equitable benchmarks for a broad variety of different quantum designs.
Hype and Misrepresentation: The subject of quantum computing is prone to unwarranted hype and conjecture, which is occasionally stoked by businesses that conceal or distort facts in order to affect stock markets and raise money. “Quantum advantage” claims have frequently been denied. Some startups have also been charged with producing “vapourware” by taking advantage of the hype surrounding quantum AI in order to make quick money.
You can also read Quantum Image Compression Advances Visual Data Processing
Applications
Quantum benchmarking is a technique for evaluating the performance of several possible uses of quantum computing rather than an application in and of itself:
Quantum Chemistry: To speed up drug discovery and material design, quantum chemistry simulates molecular and material features.
Machine Learning: Improving artificial intelligence models through innovative processing of intricate datasets.
Optimization: Resolving intricate optimization issues in supply chain management, finance, and logistics.
Materials Science: Finding novel materials with desirable qualities, such superconductors or battery catalysts, is known as materials science.
It is anticipated that these applications will cause a profound upheaval in society, revolutionizing sectors such as healthcare, banking, agriculture, energy, and the military.
Challenges
Quantum benchmarking faces several significant challenges:
Lack of Standardization: It is challenging to conduct fair comparisons throughout the industry because there isn’t a single, widely accepted criteria for quantum computer performance.
Hardware Diversity: It is difficult to develop a single benchmark that is equitable and representative of all quantum computing technologies due to the existence of several ones (such as superconducting, trapped-ion, neutral-atom, and photonic), each of which has distinct advantages and disadvantages.
Noise and Error: Benchmark results can be greatly impacted by the extreme susceptibility of quantum computers to mistakes and noise. Accurately and consistently accounting for these errors is still quite difficult.
Scalability: A lot of benchmarks created for quantum computers operating on a small scale are difficult to adapt to bigger, more intricate systems.
Segregating Speculation from Facts: Because quantum technology is still in its infancy and is so complicated, it can be challenging to tell the difference between real advancement and conjecture.
Maintaining Investor Trust: Excessive fanfare and false statements threaten investor trust, which could discourage investments and impede the advancement of the field.
History
July 2024 saw the launch of the QBI. DARPA’s Underexplored Systems for Utility-Scale Quantum Computing (US2QC) initiative is being expanded. Like QBI’s Stage C, the US2QC program, which is currently in its last stage and includes companies like PsiQuantum and Microsoft, aims to confirm and validate the creation of commercially viable quantum computers.
QBI is also associated with Quantum Benchmarking (QB), a distinct DARPA effort that focusses on establishing the “yardstick for impact” that is, what would be possible for a fully functional quantum computer to do that a conventional computer cannot. In essence, the QBI blends the hardware-validation focus of US2QC with the software and application focus of the QB program.
Companies hoping to develop a commercially viable quantum computer by 2033 were invited to a proposers day held by DARPA in September 2024 in order to obtain funding and take use of objective Independent Verification and Validation (IV&V). DARPA revealed in April 2025 that 18 businesses had been chosen for Stage A of the QBI, of which 15 had been made public and three were still negotiating contracts.
You can also read National Quantum Virtual Laboratory Gets $16 M From NSF
