leaf

Quantum computing uses quantum bits as its basic unit, leveraging principles such as quantum superposition and interference to achieve parallel computation, promising exponential speedup in solving complex computational problems. It holds significant strategic importance and scientific value, representing one of the key directions for future leaps in computational capability. Major countries around the world are continuously deepening their investments in related technology research and development, application exploration, and industrial ecosystem cultivation, leading to increasingly fierce international competition. Currently, quantum computing is at a critical stage of technological breakthroughs and application exploration. Research and development in various technological routes such as superconducting, ion traps, neutral atoms, photonic quantum, and silicon semiconductors are making continuous progress. Application explorations in industries such as finance, chemical engineering, biology, transportation, and artificial intelligence are deepening, with quantum-classical hybrid computing becoming a focal point of industry attention. Benchmark evaluation research is steadily advancing, with tech giants and startups actively developing, and domestic and international efforts to promote the construction of industrial alliances. The industry is further accelerating the construction of its ecosystem through quantum computing cloud platforms and public infrastructure, which is gradually expanding.

1. The global quantum computing landscape is continuously deepening, entering a rapid development phase.
   (1) Quantum computing is expected to bring disruptive changes and become a focal point of international competition. Quantum computing is a computational scheme that uses quantum bits as its basic unit and employs principles such as quantum superposition and interference for information processing. It possesses unparalleled information representation capabilities and super-parallel processing abilities compared to classical computing, providing exponential speedup for solving specific computationally complex problems. Quantum computing is a significant marker of the "second quantum revolution," capable of driving leaps in computational capability and expected to disrupt and reshape traditional technological systems for information processing and problem-solving, bringing unprecedented opportunities for economic and social development. Quantum computing has become a focal point for comprehensive national power competition among major countries globally, aimed at maintaining national technological sovereignty. In recent years, major technological countries have continuously strengthened their planning and layout in the field of quantum computing, with over 30 countries engaging in quantum information planning focused on quantum computing.

The United States is one of the earliest countries to conduct research in quantum computing, emphasizing government-guided advancement of quantum computing development. National strategic deployments are unfolding across multiple dimensions, including top-level design, organizational mechanisms, special plans, and ecological construction. In December 2023, the U.S. National Science and Technology Council released the "NQI 2024 Annual Report," indicating that actual investments in the quantum information field have exceeded the original five-year investment plan of $1.275 billion under the NQI Act by more than double, with cumulative investments of $3.939 billion from 2019 to 2023, and an expected investment of $968 million in 2024, with quantum computing investments accounting for the highest proportion, totaling approximately $1.4 billion over five years. European countries have been paying attention to and continuously supporting the development of quantum computing since the 1990s. In recent years, European countries have laid out and introduced a series of quantum information-related strategies and special plans, aiming to gain an advantage in the global quantum technology competition. In 2024, the European Union released a new version of the Quantum Flagship program, "Strategic Research and Industrial Agenda," proposing short-term (2027) and medium-term (2030) development goals and recommendations in four major fields, including quantum computing, through basic research, promoting industrialization, and strengthening infrastructure construction, to achieve leadership in quantum technology, industrial ecology, and key enabling factors. China places great importance on the development of the quantum information field represented by quantum computing, promoting the formation of a comprehensive research layout through the establishment of national laboratories and the implementation of major scientific and technological projects. The 2024 "Government Work Report" mentioned actively cultivating emerging industries and future industrial fields, formulating future industrial development plans, opening up new tracks in quantum technology, life sciences, and creating a number of pilot zones for future industries.

• In recent years, more than twenty provinces and cities in China have proposed planning deployments in local "14th Five-Year" scientific and technological and information technology industry development plans, focusing on quantum computing basic research, application exploration, and industrial cultivation. Additionally, countries such as the United Kingdom, Japan, Canada, India, Australia, Denmark, South Korea, Ireland, and Singapore also attach great importance to the development of quantum computing, successively releasing quantum information development strategies, focusing on top-level planning, special plans, organizational mechanisms, cutting-edge research, application exploration, industrial cultivation, and talent training to build competitive capabilities in quantum computing.

(2) Technological innovation continues to be active, gradually becoming a hotspot in frontier scientific research. Quantum computing technology innovation remains active and is gradually becoming a research hotspot in the field of frontier technology, as shown in the number of global quantum computing research papers and patents published in recent years.

The number of global quantum computing papers published has increased approximately fourfold over about ten years, reflecting the increasing activity in quantum computing research. From the publication trend, the annual growth of papers from 2013 to 2017 was relatively small, with an average annual increase of about 70 papers. However, starting in 2017, the growth rate accelerated significantly, especially from 2019 to 2021, where the annual increase exceeded 300 papers. Based on past growth trends, it is expected that the number of global quantum computing papers will continue to grow in the coming years, with related research increasing.

From 2013 to 2023, a total of 15,437 global quantum computing invention patent applications were filed, with a total of 5,417 patents granted worldwide. From the application trend, the rapid development phase began in 2013, with the annual application volume showing a rapid growth trend, peaking at 2,866 applications in 2021. The number of applications slightly declined in 2022, and the number of applications in 2023 decreased due to delays in public disclosure, but is expected to maintain an upward trend. From the granting trend, there has been a steady growth since 2013, reaching a peak of 1,384 patents granted in 2023, with the number of grants in 2024 expected to decline due to statistical timing but still maintain an upward trend throughout the year.

The statistics of the top ten countries in terms of the number of quantum computing papers reflect the research output and influence of various countries. In terms of publication volume, the United States and China occupy the top two positions, with 5,430 and 4,813 papers respectively, far ahead of other countries, reflecting the activity and leading position of both countries in quantum computing research. Germany, the United Kingdom, and Japan follow closely, with publication volumes of 1,955, 1,441, and 1,421 papers, respectively, also demonstrating strong research activity. Based on the average citation frequency per paper (i.e., the average number of citations per paper), Australia ranks first with an average of 41 citations per paper, indicating a high level of recognition and influence in its related research. The average citation frequency for the United States and Canada is 38 citations per paper, while Germany and the United Kingdom also show high influence. Although China ranks second in the number of papers, its average citation frequency is relatively low at only 19 citations, indicating that the number of high-level papers in China needs improvement.

Quantum computing includes different technological routes, with statistics on research papers in five mainstream technical directions: superconducting quantum computing, ion trap quantum computing, neutral atom quantum computing, photonic quantum computing, and silicon semiconductor quantum computing, reflecting the level of attention in different subfields of quantum computing. It can be seen that all five technological routes are receiving widespread attention, with publication volumes showing an upward trend. Among them, the publication volume of superconducting quantum computing and neutral atom quantum computing has grown particularly prominent.

The main source countries for quantum computing patent applications reflect the output and contributions of major countries/regions in quantum computing technology. The main source countries for quantum computing patents are China and the United States, accounting for 39% and 28% respectively, while Japan, Europe, and South Korea account for approximately 5%, 3%, and 2% of patent applications. This reflects the high level of technological innovation capability and activity in the field of quantum computing in the aforementioned countries/regions, with China and the United States standing out in terms of technological output and contribution.

2. Research on quantum computing technology is progressing in an orderly manner, but still faces multiple challenges.

(1) Multiple technological routes are competing, making it difficult to form a focused solution in the short term. Currently, quantum computing is characterized by the parallel development of various hardware technology routes. Different technological routes can be categorized into two types: one represented by the superconducting route and silicon semiconductor route, which are artificial particle routes, and the other represented by ion trap routes, neutral atom routes, and photonic quantum routes, which are natural particle routes. The former has advantages in scalability but is highly dependent on processing conditions for improving metrics such as logical gate fidelity and qubit control. The latter has advantages in qubit homogeneity and high logical gate precision, but faces challenges in achieving larger-scale systems. In recent years, multiple technological routes have continuously optimized key metrics such as qubit scale, quality, and decoherence time, steadily improving technological levels while maintaining a diversified and competitive development pattern, with significant uncertainty in route convergence, making it difficult to form a focused solution in the short term.

The superconducting technology route is based on superconducting Josephson structures to create two-level systems, offering advantages such as good scalability, ease of control, and compatibility with integrated circuit processes, making it one of the most closely watched and rapidly developing technology routes. In recent years, new achievements have been made in the development of superconducting quantum computing prototypes. By the end of 2023, IBM launched the 1121-qubit superconducting quantum processor Condor and the 133-qubit superconducting quantum processor Heron. In 2024, the Chinese Academy of Sciences developed a 504-qubit superconducting quantum computing chip named "Xiaohong." The Beijing Quantum Institute's joint team achieved the integration of five hundred-qubit-scale quantum chips with classical computing resources, reaching a total of 5,907 qubits. The source quantum launched a 72-qubit superconducting quantum chip named "Wukong." Numerous research achievements based on the superconducting route have emerged, with the Shenzhen Quantum Institute's joint team verifying the feasibility of using distributed quantum processors to simulate topological phases based on distributed superconducting quantum processors. Terra Quantum announced the realization of Flowermon-type superconducting qubits based on twisted copper oxide van der Waals heterostructures. A joint team from Tsinghua University simulated Fibonacci anyon braiding on superconducting quantum processors, with experimental results showing that the quantum dimension of Fibonacci anyons is very close to the theoretical golden ratio of 1.618. Overall, the superconducting quantum computing route has made rapid breakthroughs in qubit scale and quality, remaining one of the most prominent quantum computing technology routes.

The ion trap technology route uses the hyperfine or Zeeman energy levels of ions confined in radio frequency electric fields as qubit carriers, coherently manipulated through lasers or microwaves. The full connectivity of ion trap qubits gives them advantages in control precision and coherence time. In recent years, key metrics such as the number of trapped ions and logical gate control fidelity have continuously improved, with engineering technology research deepening. By the end of 2023, a joint team from Tsinghua University demonstrated various quantum error mitigation techniques for error correction in complex quantum circuits using ion trap systems. In 2024, the Quantinuum ion trap prototype Model H1 achieved single/double qubit logical gate fidelities of 99.9979% and 99.914%, respectively, with a quantum volume of 104857613, and launched a 56-qubit ion trap prototype Model H2-1, with single/double qubit logical gate fidelities of 99.997% and 99.87%. Tsinghua University achieved stable confinement and cooling of a 512-ion two-dimensional array and quantum simulation calculations with 300 ion qubits. Oxford Ionics combined ion trap technology with silicon chip technology to launch a new electronic quantum bit control technology with better scalability and lower noise characteristics. The ion trap quantum computing route faces bottleneck challenges such as large-scale qubit expansion, high integration measurement and control, and modular interconnection, and it remains to be seen whether it can stand out in future route competition.

The neutral atom technology route uses optical tweezers or optical lattices to confine atoms, with laser excitation of atoms into Rydberg states for logical gate operations or quantum simulation evolution, holding certain advantages in coherence time, control precision, and scalability. In recent years, there have been many research achievements in expanding qubit scale based on the neutral atom route. In 2024, the Technical University of Darmstadt in Germany published experimental results on controlling an array of 1,305 single-atom qubits. Infleqtion released a roadmap for atomic quantum computing, planning to launch a 1,600-qubit prototype in 2024. The UK National Quantum Computing Centre signed commercial contracts with companies such as QuEra and Infleqtion to deploy neutral atom quantum computing prototypes and build testing platforms. Pasqal announced the successful capture of approximately 1,110 atoms in 2,080 trap sites. Recent research and experiments in the neutral atom route have shown remarkable performance, with the potential for breakthroughs in quantum simulation applications, rapidly rising in the competition among multiple technological routes. The photonic quantum technology route utilizes multiple degrees of freedom of photons for encoding.

Overall, the difficulty of technological breakthroughs and the development application prospects of multiple hardware routes vary, each with its advantages and disadvantages, and they are still in a parallel development phase, with no clear conclusion on which system is optimal. Currently, the performance level of quantum computing prototypes is still far from achieving large-scale fault-tolerant universal quantum computing, and the core elements of technological breakthroughs are high-precision scalable quantum computing prototype qubit counts, which means that the design, manufacturing, and control of qubits face significant challenges, requiring continued collaborative efforts from academia and engineering in the future.

(2) Research on quantum error correction is deepening, but practical gaps remain significant. Quantum error correction is used to protect qubits from noise and other interferences, making it one of the crucial elements for quantum computers to truly realize their immense potential. The basic idea of quantum error correction schemes is to use redundant qubits to detect and correct errors in qubits, thereby restoring the original quantum state. These redundant qubits are also known as quantum error correction codes, which ensure the correctness of quantum computation even in the presence of strong environmental noise and interference. Compared to classical error correction codes, the construction of quantum error correction codes is more complex due to the inherent characteristics of quantum systems, such as the no-cloning theorem limiting the precise replication of non-orthogonal unknown quantum states, meaning that quantum error correction codes cannot use simple copying operations to increase redundancy. Since the concept of quantum error correction was proposed, various quantum error correction coding schemes using different principles have emerged, among which surface codes, as a type of two-level coding, have gained widespread attention due to their good scalability, requiring only nearest-neighbor physical qubit interactions, high fault tolerance thresholds, and applicability across multiple routes. With the continuous improvement of quantum computing hardware levels, quantum error correction research has a better physical foundation, and research continues to deepen, achieving many new developments.

In 2024, the Alice&Bob company joint team proposed a quantum error correction coding scheme based on bosonic cat state qubits and quantum low-density parity-check codes, achieving 100 high-reliability logical qubits (error rate <10−8) based on 1,500 physical qubits. A joint team from Tsinghua University proposed a quantum error correction scheme based on bosonic encoding and applied it to multiple logical qubits to achieve entanglement protection, increasing the coherence time of entangled logical qubits by 45%, and for the first time experimentally demonstrated Bell's inequality using logical qubits. IBM proposed a quantum error correction scheme based on quantum low-density parity-check codes, achieving a 0.7% error threshold, which can protect 12 logical qubits using 288 physical qubits when assuming a physical error rate of 0.1%. The Quantinuum joint team constructed four logical qubits using 30 physical qubits, reducing the error rate during entanglement of logical qubits to 10−5, nearly 800 times lower than the error rate of entangled physical qubits at 8 × 10−3. With the rapid development of quantum computing hardware performance and error correction-related control technologies in recent years, quantum error correction research and experimental validation have deepened and achieved significant progress. However, the current lowest error rate of logical qubits is still far from the practical requirements for quantum computing, and future efforts need to focus on several aspects: researching quantum error correction based on high-dimensional quantum resources, exploring distributed quantum error correction architectures, achieving quantum system control that is immune to specific noise from a theoretical perspective, building evaluation systems for practical quantum error correction schemes, and exploring operations related to fault-tolerant quantum logic gates, among others. In summary, practical quantum error correction has become one of the key research and breakthrough directions in the industry, and constructing logical qubits based on quantum error correction will be the next important milestone. To achieve this goal, continuous research and breakthroughs are still needed in the future.

(3) Quantum computing software continues to diversify, with maturity needing improvement. Quantum computing software provides developers with the necessary tools to use quantum computing hardware and run quantum algorithms, and is currently in a rapid development phase. As a structured toolset, quantum computing software needs to be developed and designed based on quantum computing principles, providing application development capabilities, compilation capabilities, hardware measurement and control capabilities, and EDA design development capabilities for different technological routes. The industry is laying out in multiple directions, and the system architecture is gradually taking shape, as shown in the following figure.

Among them, application software matches the demands of different industries and conducts demand mapping, compilation software is the foundation for achieving software development functions, measurement and control software supports the normal and efficient operation of quantum computers, and EDA software is key to enhancing the engineering level of quantum computing hardware research and manufacturing. Different quantum computing software has unique functions, and during user usage, each plays its role.

Application software provides a toolkit for creating and operating quantum programs, including algorithm libraries, development components, and debugging optimization tools, supporting developers in designing and implementing various complex quantum programs and obtaining execution results.

• In 2024, Quantinuum released version 0.4.0 of its quantum natural language processing software "lambeq," improving usability while enhancing string graph processing speed. HQS delivered quantum simulation software "HQS Noise App" to the Leibniz Supercomputing Centre, which can be used to simulate quantum mechanical systems. Microsoft Azure Quantum Elements software launched two new features for generating chemistry and density functional theory acceleration, assisting users in conducting research in chemistry and materials science. Future application development software needs to expand application scenario research, enrich the types of computational problems, improve algorithm running efficiency, and enhance support capabilities across hardware backends.

Compilation software standardizes quantum programming boundaries and achieves correct compilation and execution of quantum programs while providing a complete set of syntax rules to coordinate and constrain compilation operations.

• At the end of 2023, NVIDIA released version 23.10 of its quantum circuit simulation software cuQuantum, updating API functions and providing support for NVIDIA's Grace Hopper system. In 2024, IBM launched an updated version of Qiskit software, improving the speed of quantum hardware circuit optimization and resource usage. Intel released version 1.1 of its quantum software development kit. Quantum Circuits launched integrated quantum software for real-time management of qubit error detection during algorithm execution. Future compilation software needs to enhance software-hardware collaborative compilation capabilities while continuously updating and iterating, improving core functions such as scheduling, optimization, and debugging.

Measurement and control software is mainly used for controlling, processing, and computing quantum computing hardware, supporting measurement result feedback and chip calibration functions.

• In 2024, Isde Technology integrated Q-CTRL's Boulder Opal hardware optimization and automation features into its quantum control system to provide better quantum processor characterization and optimization functions. QuantrolOx released a quantum bit automation control software platform Quantum Edge, supporting quantum chip monitoring, workflow automation, and data visualization.

• The challenges faced by measurement and control software mainly lie in quantum error correction support capabilities, mapping capabilities between physical and logical qubits, automation, and process-oriented aspects. EDA software can provide chip design, optimization layout, simulation verification, and manufacturing testing functions for quantum computing chips.

• In 2024, Isde Technology launched the EDA simulation tool QuantumPro for superconducting quantum processor design, enabling circuit schematic design, layout construction, electromagnetic analysis, nonlinear circuit simulation, and quantum parameter extraction functions.

• EDA software needs to continuously improve in terms of functional integrity, simulation efficiency and accuracy, and optimization effects to achieve the goal of enhancing the efficiency and quality of quantum chip design. Due to the current uncertainty in the hardware technology routes of quantum computers and the lack of a fully unified general architecture, quantum computing software is still in the early stages of development and ecological construction, showing a diversified and differentiated development trend. Different types of software have varying functions, but in terms of technological maturity, stability, and user experience, they are far from the completeness of classical software. With the improvement of hardware capabilities and algorithm advancements, future quantum computing software needs to continue to advance in key areas such as quantum programming languages, algorithm libraries, quantum intermediate representations, hardware interfaces, and optimization, laying the foundation for achieving more efficient and reliable quantum computing applications in the future.

(4) Support and guarantee systems are becoming increasingly important, and indicators need further improvement. Quantum state information is easily disrupted by complex environmental noise and non-ideal characteristics within quantum systems, necessitating extremely stringent environmental support systems and high-precision measurement and control systems for quantum computers to operate. The quantum computing support and guarantee system is a crucial component of technology research and prototype engineering development, primarily including environmental equipment, measurement and control systems, and key equipment components, with varying bottleneck challenges faced by different parts. Environmental equipment is a necessary support component and infrastructure for ensuring the stable operation of quantum computers, including ultra-high-power dilution refrigerators, GM pulse tube refrigerators, ultra-high vacuum chambers, and pump groups.

In 2024, Bluefors launched a super-compact LD dilution refrigeration system. In recent years, China has also made significant achievements in devices such as dilution refrigerators, with the Guoshield Quantum ezQfridge dilution refrigerator completing delivery testing, and Source Quantum launching its self-developed Source SL1000 dilution refrigerator. The environmental equipment required for quantum computers operating on different technology routes varies, and future breakthroughs are still needed in terms of technological levels, core indicators, equipment engineering, miniaturization, and integration to support the large-scale expansion of qubit numbers in quantum computers. Precise quantum control technology and efficient readout technology are crucial for achieving reliable single-qubit operations.

3. The exploration of quantum computing applications continues to gain momentum, with practical implementation awaiting breakthroughs.

(1) Multi-field application exploration is advancing, but practical implementation still needs to accelerate breakthroughs. Currently, quantum computing is at a critical stage of transitioning from frontier research to practical implementation, with widespread and active multi-party application exploration being key to driving quantum computing technology toward application. The industry is actively seeking specific application scenarios that match industry needs, aiming to provide services for various industry application fields in the future. Typical application areas include finance, chemical engineering, biology, transportation, and artificial intelligence.

According to the "Quantum Technology Monitoring" research report released by McKinsey in 2024, quantum computing is expected to accelerate development in the next five to ten years, with the market size potentially reaching trillions of dollars by 2035.

The financial sector has numerous potential application scenarios for quantum computing, including financial risk management, portfolio analysis, simulated quantitative trading, and financial market forecasting. At the end of 2023, Multiverse Computing and Moody's jointly launched the QFStudio platform, providing quantum computing solutions for financial domain application exploration. In 2024, the Chicago Quantum Exchange released a report suggesting that quantum computing is expected to shorten the time to obtain optimal solutions and improve prediction accuracy in the financial sector. Citibank and Classiq are jointly researching quantum solutions for portfolio optimization based on the Amazon Braket platform, constructing a performance-optimized portfolio based on expected returns and risk levels. In the chemical engineering field, quantum computing applications can be used to simulate chemical molecular structures, chemical reactions, and design chemicals more efficiently and with lower energy consumption. In 2024, BP and ORCA used hybrid quantum-classical machine learning methods to model molecular conformations, exploring the potential of quantum computing to enhance the performance of machine learning algorithms in the chemical field. Microsoft collaborated with the U.S. Department of Energy's Pacific Northwest National Laboratory to use quantum computing to filter new battery materials down to a few candidates, with experimental results indicating a significant reduction in screening time. The Quebec Hydro Company is exploring solutions to complex energy problems using quantum computing, aimed at predicting energy demand and designing and operating sustainable energy systems. In the biological field, quantum computing applications mainly focus on early disease diagnosis, drug research and screening, drug testing, genomic data research, and protein structure prediction.

In 2024, Boehringer Ingelheim's quantum laboratory published a paper discussing the current state of quantum computing applications in drug discovery, suggesting that quantum computing is expected to produce practical applications in drug design in the future. IBM and the Cleveland Clinic collaborated to use quantum-classical hybrid methods to predict protein structures, effectively improving prediction accuracy. Novonesis and Kvantify collaborated to demonstrate the enzymatic reaction calculations of carbonic anhydrase using hybrid quantum-classical computing methods, which may aid in biological process research and industrial carbon dioxide capture. In the transportation sector, quantum computing applications can be used for traffic flow optimization algorithms and real-time predictions, as well as dynamic path planning.

In 2024, IonQ collaborated with the German Center for Basic Research to apply quantum computing to optimize flight boarding gates, shortening passenger transfer times and aircraft docking times while improving boarding gate service efficiency. Pasqal and Thales demonstrated the potential of quantum computing in solving satellite planning problems based on neutral atom quantum processors. The Singapore Quantum Technology Center used 8 and 13 qubits to solve vehicle routing problems of 128 and 3,964 routes, improving the efficiency of solving combinatorial optimization problems.

(2) The number of quantum computing cloud platform providers is gradually increasing, but functions generally need strengthening. Currently, quantum computers have high thresholds for hardware and software usage, strict hardware environment requirements, and high operational costs, making it difficult for enterprises and individual users to deploy locally. Against this backdrop, quantum computing cloud platforms have emerged, integrating quantum computing with classical cloud services to provide users with remote access to quantum computers via the internet. Quantum computing cloud platforms, with their flexible service models, convenient access methods, and rich application scenarios, are gradually becoming one of the important development directions for quantum computing and are expected to become the main application form for providing quantum computing services in the future. Dozens of quantum computing cloud platforms have emerged globally, with typical cloud platforms showing a thriving trend.

Currently, the quantum computing processors that quantum computing cloud platforms can provide include superconducting, ion trap, neutral atom, photonic quantum, and silicon semiconductor technology routes. The access modes for backend hardware of quantum computing cloud platforms can be mainly divided into three categories. The first category is the self-developed equipment access mode, where cloud platform providers have the capability to independently develop quantum computing hardware, providing self-developed quantum computers or quantum simulators based on classical computing on the cloud platform. Representative enterprises or institutions include IBM, IonQ, Xanadu, Rigetti, Source Quantum, Guoshield Quantum, and Beijing Quantum Institute. The second category is the cloud service access mode, where cloud platform providers leverage their cloud service capabilities to access hardware and software from other suppliers on the cloud platform. Representative enterprises or institutions include Microsoft, Amazon, Strangeworks, Arc Quantum, China Mobile, and China Telecom. The third category is the hybrid access mode, which is a combination of the above two access modes, meaning that while accessing self-developed hardware, it also supports calling hardware resources from other suppliers. For example, the IBM Quantum Cloud platform can access self-developed quantum processors as well as hardware resources from suppliers such as Rigetti, Xanadu, AQT, and IonQ.

Internationally, tech giants such as IBM, Google, and Microsoft, as well as startups like IonQ, Xanadu, and Rigetti, are actively laying out quantum computing cloud platforms, attracting a large number of developers, researchers, and enterprise users by providing quantum computing processors, simulators, and development tools.

At the end of 2023, IBM integrated Q-CTRL's error suppression technology software Q-CTRL Embedded into its cloud platform, with tests showing that the complexity of quantum algorithms that can be run increased by tenfold and the success rate improved by approximately 1,000 times after error suppression. IonQ provides the Forte quantum computer on the Amazon Braket platform. Amazon launched the "Braket Direct" program on the Amazon Braket cloud platform, allowing users to reserve the computing power of specific quantum processors within a set time period without waiting in line. In 2024, AQT collaborated with Deutsche Telekom to provide users with cloud access to its quantum computer. Domestically, quantum computing companies such as Source Quantum, Guoshield Quantum, and Arc Quantum, as well as telecom operators like China Mobile and China Telecom, have successively launched quantum computing cloud platforms, indicating that quantum computing companies attach great importance to the development of cloud platforms and reflecting that telecom operators recognize the potential value of quantum computing in enhancing network performance and strengthening secure communication, striving to jointly promote the application and industrialization process of quantum computing.

At the end of 2023, the Cloud Capability Center of China Mobile and Boson Quantum jointly launched the "Wuyue Quantum Computing Cloud Platform - Hengshan Photonic Quantum Computing Power Platform." The "Wuyue" quantum computing cloud platform of China Mobile lays out multi-format quantum computing power networking, multi-mode quantum algorithm program design, and diversified quantum scenario algorithms, aiming to expand the application boundaries of quantum computing. In 2024, the Beijing Quantum Institute, in collaboration with the Institute of Physics of the Chinese Academy of Sciences and Tsinghua University, released the Quafu quantum cloud computing cluster, which provides resources for five hundred-qubit-scale quantum chips and integrates classical computing resources. The Quantum Information and Quantum Technology Innovation Research Institute of the Chinese Academy of Sciences developed and delivered the 504-qubit quantum computing chip "Xiaohong," with plans to open it to the world through the "Tianyan" quantum computing cloud platform of China Telecom. Qike Quantum launched a quantum-classical hybrid computing cloud platform "<Qu|Cloud>", providing access to a 20-qubit ion trap quantum computing processor and a CPU/GPU-based quantum computing simulator, supporting multiple programming modes and algorithm libraries.

Overall, domestic quantum computing cloud platforms still have a significant gap compared to international advanced levels in terms of cloud platform functions, application exploration, business model, and user influence, and further improvements are still needed in the future. Quantum computing cloud platforms have become one of the important tools for users to access quantum computing resources, conduct experimental validation, and explore applications. With the continuous advancement of quantum computing technology and the increasing maturity of cloud platform functions, future quantum computing cloud platforms will show three development trends: first, innovation and expansion of service models, evolving from current infrastructure services to richer platform services and application services; second, deep integration and collaboration across platforms and industries, promoting multi-field quantum computing applications and implementation; third, the construction of intelligent and automated operation management and security protection systems, enhancing user experience and data security levels. The development of quantum computing cloud platforms requires joint efforts from the industry in multiple directions. First, continue to increase R&D investment to enhance the maturity and stability of quantum computing technology, thereby supporting the long-term stable operation of quantum computing cloud platforms; second, strengthen the construction of data security and privacy protection mechanisms to ensure the security and controllability of user data; finally, promote the development of standardization and interoperability to lower the barriers to interaction and use between different platforms, facilitating the popularization and application of quantum computing.

(3) Benchmark evaluation research is steadily advancing, with results and challenges coexisting. With the development of quantum computing prototypes and the exploration of applications, benchmark evaluation has gradually gained attention, and how to accurately and efficiently assess the performance of quantum computing systems has become a focal point of industry concern, providing important references for users to analyze the development level of the quantum computing technology industry. Quantum computing benchmark evaluation is a key technology for characterizing hardware performance indicators and evaluating system capabilities, which not only helps promote the development and application of underlying quantum computing hardware but also serves as a crucial bridge connecting theoretical research and practical applications.

The development of quantum computing benchmark evaluation has been rapid, with the industry proposing a series of evaluation benchmark methods. These benchmark methods typically include various tasks with specific functions, such as fidelity testing of quantum gate operations, coherence time assessment of qubits, and execution efficiency of quantum algorithms, aiming to provide relatively fair comparison means for different quantum computing systems, helping researchers gain a more comprehensive understanding of system performance. The framework of the quantum computing benchmark evaluation system can be divided into qubit level, quantum circuit level, system level, algorithm level, and application level, with different characteristics and focuses presented at each level. The lower benchmarks, such as qubit level and quantum circuit level, have a high correlation with hardware and can fully reflect the differences between various technological routes. The parameters and indicators at the lower level are relatively more dispersed and specific, making it easier for researchers familiar with technical details to accurately identify problems and propose solutions. As the levels rise, for example, the system level...

In recent years, the industry has actively carried out research on quantum computing benchmark evaluation, aiming to evaluate the comprehensive performance of quantum computing systems using more objective methods. At the end of 2023, IBM proposed the Every Layer Gate Error (EPLG), which can more accurately assess crosstalk and can also be used to estimate the number of circuits required for error mitigation, while updating the definition of Circuit Layer Operations per Second (CLOPSh) to more realistically reflect hardware performance. EPLG, CLOPSh, and the quantum volume (QV) first proposed by IBM can comprehensively evaluate the performance of quantum computing systems from the dimensions of scale, quality, and speed. In 2024, QED-C updated the application-oriented evaluation benchmark suite, expanding evaluation benchmarks for algorithms such as HHL, VQE, and quantum machine learning, and introducing parameters such as the quality of computational results (e.g., final ground state energy, classification accuracy) and computational costs for quantum computing performance evaluation. The U.S. DARPA launched a new quantum benchmark testing program (QBI), primarily targeting benchmark testing for quantum computing algorithms and applications, and assessing the feasibility of building industrial-grade quantum computers.

As quantum computing technology continues to develop, various testing benchmark studies have become particularly important. However, quantum computing benchmark evaluation research also faces a series of challenges, such as the objectivity and fairness of benchmarks being a major concern in the industry. In 2024, Quantinuum pointed out in its report that the #AQ benchmark may lead to an overestimation of quantum computer performance in certain cases, primarily due to the application of error mitigation techniques and circuit compilation strategies, which can enhance efficiency and accuracy in specific usage scenarios but may mislead overall performance assessments. Therefore, when evaluating and comparing the performance of different quantum computers, researchers must consider these factors to ensure the objectivity and fairness of evaluation results. Quantum computing benchmark evaluation research plays a crucial role in assessing development status, promoting industry development, and connecting theory with practical applications. Currently, research on quantum computing benchmark evaluation is continuously deepening both domestically and internationally, achieving results while also facing numerous challenges. In the future, the industry needs to continuously improve the evaluation system, update evaluation schemes, and establish evaluation standards to more accurately and comprehensively showcase the actual performance of quantum computers, promoting continuous progress in the industry.

(4) Quantum-classical integration has become a focal point, and the architecture of the technical system is crucial. The quantum computing technology industry is currently in a phase of vigorous development; however, the operation and maintenance of current quantum computers still face significant challenges. Future large-scale commercialization must bridge the gap between theoretical advantages and practical application value. The industry is gradually realizing that neither pure quantum computing nor classical computing can meet all computational needs, thus necessitating an organic integration of the two to form more powerful computational capabilities. In this context, quantum-classical hybrid computing combines quantum computing and classical computing, fully leveraging the advantages of both to jointly solve complex problems.

As a new computing model, quantum-classical hybrid computing has two fundamental characteristics: hybridization and collaboration. Hybridization refers to the simultaneous inclusion of quantum computing and classical computing within a single system, forming a heterogeneous computing model. Quantum computers can be divided into universal gate-type quantum computers and specialized quantum computers. Universal gate-type quantum computers currently exist in various technological routes, including superconducting, ion trap, neutral atom, photonic quantum, and silicon semiconductor, with significant differences in technical principles, performance indicators, and maturity among different routes. Specialized quantum computers mainly include quantum annealers and coherent Ising machines. Classical processors mainly include central processing units (CPUs) and graphics processing units (GPUs). Heterogeneous computing integration includes both the hybridization between universal gate-type quantum computers and specialized quantum computers, as well as the hybridization of various quantum computing architectures with various classical computing architectures. Collaboration refers to the quantum computer's responsibility for processing quantum information, such as quantum state preparation and measurement, while the classical computer handles classical information, such as logical operations, floating-point operations, algorithm analysis, and optimization. By designing algorithms and interfaces, the quantum computing part can collaborate with the classical computing part to jointly complete computational tasks. Quantum computers are suitable for solving problems involving parallel computation, matrix operations, and linear algebra, while classical computing excels at logical operations, floating-point operations, and has relatively complete programming development tools, operating systems, and algorithm libraries. The core idea of quantum-classical hybrid computing is to leverage the advantages of quantum computing to accelerate the resolution of specific problems while utilizing the stability and ease of use of classical computing to ensure the accuracy and reliability of computations.

A preliminary proposed architecture for quantum-classical hybrid computing technology systems can be divided into seven layers: application layer, development tools layer, algorithm layer, programming framework layer, task scheduling layer, resource management layer, and physical resources layer. The application layer includes typical application fields for quantum-classical hybrid computing, including quantitative finance, energy materials, biomedicine, transportation logistics, and information communication, primarily providing computing services to industry users through packaged software, functions, or custom-developed forms. The development tools layer provides development and debugging tools for quantum-classical hybrid algorithms, including Jupyter Notebook, WebIDE, etc. The algorithm layer provides typical quantum-classical hybrid algorithms for the application layer, with representative algorithms including Variational Quantum Eigensolver (VQE), Quantum Approximate Optimization Algorithm (QAOA), Quantum Machine Learning (QML), and Quantum Neural Networks (QNN). The programming framework layer provides basic programming languages and compilation tools for algorithm development, offering interconnection interfaces for underlying hardware and upper-layer application software, while decomposing and enabling interoperability between quantum and classical computing tasks, ultimately converting high-level programming languages into hardware instruction sets to be passed to the underlying hardware. The task scheduling layer schedules the decomposed quantum and classical tasks and achieves collaboration between various quantum and classical heterogeneous computing resources. Currently, there are two main task scheduling methods: heterogeneous parallel scheduling and remote parallel scheduling, with the former achieving low-latency communication between quantum and classical systems, while the latter is relatively easier to implement. The resource management layer implements registration, monitoring, scheduling, and other functions for various physical machines, virtual machines, Docker containers, and topologies. The physical resources layer is the lowest layer, divided into classical resources and quantum resources, with classical resources including various classical computing, storage, and transportation infrastructure, and quantum resources including universal gate-type quantum computers, specialized quantum computers, and quantum circuit simulators from various technological routes.

With continuous technological breakthroughs, tech companies are gradually recognizing the importance of quantum-classical hybrid computing and are competing to lay out related research. Internationally, NVIDIA released the GPU-accelerated quantum computing system NVIDIA DGX Quantum, which is based on the NVIDIA Grace Hopper architecture superchip and the open-source quantum-classical hybrid programming model CUDA Quantum, reducing communication latency between GPUs and QPUs to sub-microsecond levels. Microsoft proposed four quantum-classical hybrid modes: batch quantum computing, interactive quantum computing, integrated quantum computing, and distributed quantum computing, gradually transitioning from remote parallel scheduling to heterogeneous parallel modes. Amazon launched the Braket Hybrid Jobs tool, achieving fully managed orchestration of quantum-classical hybrid algorithms, combining access to classical computing resources and quantum processors while supporting parameterized compilation of quantum circuits to optimize the execution process of quantum-classical hybrid algorithms. IBM's roadmap indicates that it expects to demonstrate supercomputing with quantum computing centers by 2025, integrating quantum processors, classical processors, quantum communication networks, and classical networks.

Domestically, Zhongwei Daxin launched a hybrid computing measurement and control unit suitable for classical computers and multi-path quantum measurement and control, achieving millisecond-level invocation delays between quantum-classical measurement and control instructions via PCIe interfaces. Source Quantum released a quantum-cloud hybrid solution architecture, allowing quantum computers to connect remotely with classical supercomputers via public networks, and implementing quantum-classical interaction protocols between the quantum operating system and supercomputing management scheduling modules for collaborative computing. China Telecom's quantum "Tianyan" quantum computing cloud platform offers both batch and interactive quantum-classical hybrid modes, achieving remote parallel scheduling. In the future, as quantum computing technology continues to advance and classical computer performance improves, quantum-classical hybrid computing is expected to become one of the important trends driving the computing industry forward, with both forming complementary advantages and being key to promoting technological development. Overall, the quantum-classical hybrid field will further explore application scenarios while continuously improving scheduling mechanisms, gradually establishing an industrial ecosystem step by step. Hardware manufacturers need to develop high-performance, highly stable quantum-classical hybrid computing systems to provide a strong computational foundation for the entire ecosystem; software developers need to develop efficient and user-friendly programming tools and software platforms tailored to the characteristics and needs of hybrid computing, reducing development difficulty and improving development efficiency; application service providers will leverage the advantages of quantum-classical hybrid computing to provide customized solutions for various industries, promoting the digital transformation and upgrading of industries.

4. The cultivation of the quantum computing industry is being pursued by multiple parties, and the ecosystem is gradually emerging.

(1) The industrial ecosystem is initially forming, but key links still need to be promoted. With the development of quantum computing prototype manufacturing, software development, application exploration, and cloud platform construction, upstream and downstream enterprises are continuously emerging, injecting strong momentum into the development of the quantum computing technology industry. The cultivation of the quantum computing industrial ecosystem is steadily underway, as shown in the following figure, with an increasing number of participants in each link, although the overall ecosystem is still in its infancy, and key links still need to be promoted.

The upstream of the industrial ecosystem includes environmental support systems, measurement and control systems, and core equipment components, involving many aspects such as dilution refrigerators, vacuum systems, low-temperature components, and optical devices, serving as the foundational base of the entire quantum computing industrial ecosystem. Due to the complexity of quantum computing technology, the parallel advancement of multiple routes, and the uncertainty in development trends, the upstream of the industrial ecosystem currently presents characteristics of decentralization and diversification. On one hand, decentralization increases the difficulty for suppliers to concentrate on technological breakthroughs, but on the other hand, diversification may help reduce risks such as supply monopolies that could arise from a single supplier. In comparison, Europe and the United States have a greater number of enterprises in the upstream of the quantum computing industrial ecosystem, with higher development levels, having accumulated relatively superior conditions and resources in product development, technological innovation, and market demand. Upstream enterprises in China have developed rapidly in recent years, launching various self-developed products, but there is still significant room for improvement in performance indicators, manufacturing costs, and market recognition of some key equipment components, necessitating further enhancement of product technology levels through independent research and development in the future.

Midstream enterprises in the industrial ecosystem include quantum computing prototype manufacturers and software suppliers, forming the core link of the quantum computing industrial ecosystem, and are also the part with a relatively concentrated number of enterprises. Among the global enterprises engaged in the development of quantum computing prototypes, those focusing on the superconducting route account for the largest number, exceeding one-third of the total, followed by ion traps, neutral atoms, photonic quantum, and silicon semiconductors. In terms of software, many enterprises are dedicated to building their own quantum computing software while also constructing open-source software communities to promote the development and application exploration of quantum computing technology. In comparison, most countries are laying out multiple technological routes in parallel, with European and American enterprises holding certain advantages in terms of quantity, prototype development capabilities, software development, and open-source community construction. China has laid out in several mainstream technological routes and has seen the emergence of a number of quantum computing software companies in recent years, but overall, there are still gaps compared to Europe and the United States in terms of enterprise investment intensity, output results, and innovation capabilities.

Downstream enterprises in the industrial ecosystem include quantum computing cloud platform providers and industry application enterprises, playing a crucial role in the ecosystem and being closest to users. In terms of cloud platforms, they provide cloud access for various users, sharing quantum computing resources and promoting early layout and healthy cultivation of the quantum computing industry. In terms of industry applications, users in industries such as finance, chemical engineering, medicine, and transportation are focusing on the application potential of quantum computing, opening application scenarios, and conducting application explorations, striving to find solutions to industry-specific challenges. In comparison, quantum computing cloud platforms from foreign tech giants such as IBM, Amazon, and Microsoft are leading the world in terms of resource sharing, hardware diversity, richness of application cases, and commercialization of service models. Quantum computing enterprises are actively collaborating with enterprises from different sectors to jointly explore applications of quantum computing in key industry fields. Domestic quantum computing cloud platform providers still need to improve in terms of collaborative cooperation between platforms, backend hardware levels, and exploration of business models. Traditional industry enterprises in China still need to further strengthen and improve their investment intensity, attention levels, and cooperation mechanisms with quantum computing enterprises, and in the future, proactive application exploration is needed to enhance collaborative innovation capabilities.

Quantum information technology, represented by quantum computing, has become one of the important focal points for future industrial layout. The industrial foundational capabilities support the layout and development of future industries, and comparative analysis of the industrial foundational capabilities of different countries can provide perspectives and tools for evaluating a country's comprehensive strength and international competitiveness in the field of quantum computing technology. This report builds an analysis method for the foundational capabilities of the quantum computing industry based on dimensions such as scientific research foundation, government support, commercial activities, and technological achievements.

Quantum computing: disruptive innovation in the computing power industry, the sharp spear of future technology. Quantum computers utilize the principle of quantum superposition states to achieve exponential growth in the amount of information processed. For example, using Shor's algorithm, it can crack a 2048-bit RSA password in 8 hours on 200,000 physical qubits with an error rate of 0.1%, while a classical computer would take hundreds of years to do so. From the industrial chain perspective, quantum computing chips, dilution refrigerators, and room temperature measurement and control systems have become the main components of quantum computers. According to an ICV report, the global quantum computing industry scale reached $4.7 billion in 2023, with an average annual growth rate (CAGR) of 44.8% from 2023 to 2028, expected to achieve rapid growth. Key players to watch include quantum computing overall solution providers like Guoshield Quantum and quantum measurement and control system providers like Puyuan Precision Electronics.

Quantum communication: quantum technology achieves key distribution, a solid shield for information security. The key distribution and digital signature technology based on traditional RSA algorithms face significant security risks in the era of quantum computing. Quantum secure communication converts classical keys into quantum forms, utilizing the physical characteristics of quantum non-clonability and entanglement to ensure absolute security during the key distribution process. In terms of the industrial chain, quantum key distribution devices (QKD) have become the core equipment of the industry, with upstream components including chips, light sources, single-photon detectors, and quantum random number generators, while downstream applications are primarily in government, finance, and critical infrastructure sectors. In terms of construction progress, China has formed a three-step development strategy of backbone network, metropolitan area network, and integrated space-ground network, having built a wide-area quantum secure communication backbone network exceeding 10,000 kilometers in length, with future metropolitan area network and integrated space-ground network construction expected to accelerate. Key players to watch include QKD device vendors like Guoshield Quantum and system integrators like Shenzhou Information.

Post-quantum cryptography: underlying innovation in cryptographic principles, a new solution to counter quantum attacks. Post-quantum cryptography (PQC) is a new generation of cryptographic algorithms that can resist attacks from quantum computers on existing cryptographic algorithms. The goal is to design new cryptographic algorithms that maintain high security in the face of quantum computing threats. The mainstream post-quantum cryptographic algorithms can be roughly divided into five categories based on the underlying hard problems they rely on: (1) Lattice-based PQC algorithms, (2) Hash-based PQC algorithms, (3) Code-based PQC algorithms, (4) Multivariate-based PQC algorithms, and (5) Isogeny-based PQC algorithms.

1. Lattice-based algorithms rely on the hardness of problems such as the Shortest Vector Problem (SVP) and the Closest Vector Problem (CVP) in lattices. They are considered the most promising PQC algorithms due to their balance of security, public/private key size, and computational speed.
2. Hash-based algorithms are limited to digital signatures and rely on the security properties of hash functions. They are theoretically strong but have drawbacks such as large signature sizes and limited signature counts.
3. Code-based algorithms, such as the McEliece scheme, utilize the difficulty of decoding random linear codes and are considered competitive in the PQC space.
4. Multivariate-based algorithms use systems of multivariate quadratic polynomials and are suitable for applications where algorithm efficiency is prioritized over bandwidth.
5. Isogeny-based algorithms focus on the problem of finding isogenies between elliptic curves and have shown promise but face challenges in efficiency.

The global landscape for post-quantum cryptography is actively evolving, with various countries and organizations pushing for standardization. The U.S. National Institute of Standards and Technology (NIST) is leading the charge, with a focus on selecting algorithms that can withstand quantum attacks. As quantum computing technology continues to advance, the migration to post-quantum cryptography is becoming increasingly important to ensure the security of existing cryptographic systems.

In summary, quantum computing and its associated technologies are rapidly evolving, with significant implications for various industries and applications. The ongoing research and development in quantum computing, quantum communication, and post-quantum cryptography are paving the way for a future where quantum technologies play a crucial role in shaping the landscape of computing and information security.