Quantum Computing: AI's New Engine Revolutionizing Industries

Under the premise of Moore's Law, the evolution of silicon-based integrated circuit technology is approaching its physical limits. We believe that quantum computing is expected to become one of the important solutions for a leap in computing power in the post-Moore era.

The market is still in the early stages of exploration, with overseas leaders mainly being large factories such as IBM, Google, and NVIDIA. At the same time, startups such as IONQ, Rigetti, and D-Wave have also emerged through SPAC acquisition methods. Domestic quantum startups mainly originate from university laboratories, with representative companies including Guo Dun Quantum, Ben Yuan Quantum, and Guo Yi Quantum. In terms of future applications, the chemical, financial, and pharmaceutical industries are expected to be the first to land.

As early as 2016, quantum computing was included in the "13th Five-Year Plan," and its strategic position has been relatively clear. The "14th Five-Year Plan" released in 2021 once again mentioned the innovation of quantum computing technology, further providing policy guarantees for the scientific research of quantum computing. Since 2023, with the emergence of some quantum computing innovation achievements, the timing for the industrialization of quantum computing has become increasingly mature.

Advertisement

In 2024, quantum computing policies began to be implemented intensively. In addition to promoting technological research and development, they also paid more attention to the promotion of industrialization. In January 2024, seven departments including the Ministry of Industry and Information Technology issued the "Implementation Opinions on Promoting the Innovative Development of Future Industries," proposing to strengthen the research and development of quantum computing technology, promote the coordinated development of quantum software and quantum cloud platforms, explore the application of quantum computing in vertical industries, and further provide policy guidance for the formation and improvement of the quantum computing industry chain. In March 2024, the government work report mentioned the key layout of emerging fields such as quantum information. Under multiple guidance, local governments have also introduced relevant policies on technological research and development and even industrial pilots, aiming to promote the development and landing of quantum computing.

01

Why should we pay attention to quantum computing?

The essence of quantum computing is to achieve exponential acceleration of parallel computing and form quantum superiority. Quantum computing uses quantum bits as the basic unit, using principles such as quantum superposition and interference to achieve parallel computing, which can provide exponential acceleration on some computationally difficult problems. Quantum computers have significant "quantum superiority" in terms of computational power growth compared to classical computers. Specifically, based on the principle of quantum mechanics superposition, a quantum bit can be in a coherent superposition of both 0 and 1 states at the same time, enabling the quantum processing unit (QPU) to perform parallel computing on 2^n superimposed numbers at the same time, while CPUs and GPUs can only perform calculations on n and n^2 numbers respectively. The performance of the QPU is usually represented by the number of quantum bits it contains, and currently, companies and academia are using various technologies to manufacture quantum bits inside the QPU.

With the rise of machine learning, deep learning, and big data, higher requirements have been proposed for chip technology with low power consumption, small size, and ultra-high computing speed. However, under the limitations of Moore's Law, the evolution of silicon-based integrated circuit technology has approached its physical limit. We believe that quantum computing is expected to become one of the important solutions for a leap in computing power in the post-Moore era.

The international academic community generally believes that quantum computing is expected to be commercially available in the next 5-10 years. With the launch of IBM's superconducting quantum chip Condor in December 2023, quantum computing has achieved a breakthrough of more than 1000 quantum bits. According to ICV estimates, the quantum computing industry is expected to enter a period of rapid growth after 2027, and the global scale is expected to reach hundreds of billions of dollars by 2030. As for the specific critical point, it still needs continuous tracking. However, it is more certain that quantum computing is expected to help artificial intelligence achieve a significant increase in computational power.

Under continuous exploration, the business models of domestic and foreign quantum computing-related companies have taken shape. We suggest that the current focus should be on quantum computing chip companies and other producers of core components of quantum computers, including foreign companies such as IBM, Google, IONQ, Rigetti, D-Wave, etc., and domestic companies including Guo Dun Quantum, Ben Yuan Quantum, Guo Yi Quantum, etc.02

Technical Routes of Quantum Computing: Superconducting and Ion Trap Technologies Take the Lead

The development of quantum computing hardware is currently in a stage of parallel development and open competition among various technical routes. At present, considering factors such as system scalability and control precision, superconducting systems and ion traps are in the leading position, while the advantages of neutral atom quantum computing routes in quantum simulation are becoming increasingly apparent.

Superconducting systems are currently known as the best scalable solution. Companies such as IBM, Google, Rigetti, and Origin Quantum have chosen this technical route. IBM's 1121-qubit superconducting quantum chip, Condor, released in December 2023, is the highest bit-count superconducting quantum chip publicly released to date; according to its latest quantum computing roadmap released in December 2023, IBM plans to achieve useful computing within ten years, and by 2033, quantum-centric supercomputers will include 1000 logical qubits, fully unleashing the power of quantum computing.

Ion traps are the earliest physical systems attempted to implement quantum computing, providing stable and well-isolated quantum systems, but with limited scalability. Companies such as IONQ, Honeywell, and QuantumCTek have chosen this technical route. IONQ's Forte, the world's first software-configurable quantum computer, with 32 qubits and 29 algorithmic qubits (AQ), represents IONQ's largest single-core quantum processor, and the product is used by companies like Hyundai Motor and Airbus. In addition, Quantinuum, a subsidiary of Honeywell, also launched the second-generation quantum computer H2 in May 2023, and used it to create non-Abelian anyons, taking a key step towards building fault-tolerant quantum computers.

Neutral atom technology is also emerging. This platform uses lasers to cool and control atoms to improve the scalability and fidelity of qubits; moreover, neutral atoms may have better robustness. Companies such as Atom Computing and PASQAL have chosen this technical route. Atom Computing created a 1225-site atomic array (currently filled with 1180 qubits) in its quantum computing platform in October 2023 and plans to launch this neutral atom quantum computer in 2024.

03

Huge Investment, How to Convert Technology into Demand is the Core

Quantum computing requires a large amount of investment, with at least $16.4 billion needed globally by 2027. From 2018 to 2022, global quantum computing financing grew significantly, with a compound annual growth rate (CAGR) of 140.6%. In 2023, investment fell, mainly due to reduced venture capital under market fluctuations, and the market is more focused on targets with commercial feasibility and profit paths. According to the prediction of the Global Frontier Technology Consulting Institution (ICV), before 2027, the NISQ era industry needs at least $16.4 billion in cumulative investment in the field of quantum computing. At present, the long waiting period for technological breakthroughs and the scarcity of paying users will pose a continuous challenge to the quantum computing industry. How to convert advanced technology into effective demand and obtain continuous patient funding is the primary issue that the quantum computing industry needs to focus on next.

Special Purpose Acquisition Companies (SPACs) are increasingly becoming the main way for quantum companies to go public. For example, in February 2022, D-Wave went public on the New York Stock Exchange through a SPAC, raising $340 million, with a valuation of about $1.6 billion at the time. From 2018 to 2022, U.S. quantum companies raised $2.36 billion in financing, of which IONQ raised $650 million through a SPAC on the New York Stock Exchange in September 2021, setting a new record for quantum technology company financing; Chinese companies raised a total of 335 million yuan, all from venture capital, with some undisclosed amounts resulting in an underestimation of the total amount.At present, the industry landscape of quantum computing is led by American technology companies, with Chinese companies actively catching up. According to IPRdaily statistics, as of October 18, 2022, the top 100 enterprises in global quantum computing technology invention patents mainly come from 18 countries and regions, with the United States accounting for 40%, China 15%, and Japan 11%. Among them, IBM from the United States tops the list with 1323 patents, followed by Google and Canadian quantum computing company D-Wave with 762 and 501 patents respectively, ranking second and third; while Chinese company Origin Quantum ranks sixth on this list with 234 patents.

With the gradual advancement of research and development work on various routes of quantum computers, the selection of upstream hardware equipment and components required for the whole machine is becoming increasingly clear. At the same time, the software systems of quantum computers are also continuously catching up, and the composition of various links in the entire industry chain is gradually becoming clear and improving, with more and more participants in each link. Specifically, the upstream includes hardware such as dilution refrigerators, measurement and control systems, low-temperature components, vacuum systems, lasers, optical detectors, and software development tool kits; the midstream industry includes whole machines, system software, application software, etc.; the downstream includes quantum cloud platforms and various application cooperations in defense, military, aerospace, finance, medicine, automotive, transportation, chemical, material, etc.

06

Future Development Trends: Hybrid Quantum Computing May Become a New Trend in the Future

In the next 3-5 years, quantum computers may have some limitations, such as not being as good as CPUs in organizing and accessing memory in QPU, and not being as good as GPUs in rendering complex graphics. Therefore, classical systems may be responsible for handling tasks such as data preparation, visualization, and error correction, while quantum systems focus on handling complex calculations. This hybrid quantum computing architecture may become a new trend in the future.

In the industry, some major companies have already begun to explore and invest in the development of hybrid quantum computing. IBM believes that the future of computing is centered on quantum supercomputers, where QPU, CPU, and GPU all work together to accelerate computing; NVIDIA launched an open platform for hybrid quantum system programming, QODA, in 2022.

From a competitive perspective, IBM, Google, IONQ, Rigetti, and Quantinuum, which are currently deploying superconducting technology and ion trap technology, are in a leading position in the industry, and NVIDIA's CUDA Quantum research and development is progressing rapidly.

From a technical classification or industrial chain layout perspective, superconducting and ion trap technologies continue to lead, and quantum system layout companies may reshape the quantum industry ecosystem. Looking at typical companies in various fields, the integration of large-scale superconducting quantum bits in superconducting technology is expected to accelerate; in the field of ion trap technology, IONQ, as the industry leader in ion trap technology, has made certain commercial progress; in the field of neutral atom technology, Atom Computing announced in October 2023 that it is expected to launch a 1225-qubit neutral atom quantum computer in 2024, which is expected to accelerate the catch-up of neutral atom technology; in the layout of quantum computing systems, NVIDIA is working with its partners on cooperation in various aspects including software and hardware based on CUDA QUANTUM, and may reshape the quantum industry chain ecosystem in the future.

The industry is still in its early stages, but quantum computing companies and scientific research institutes are actively exploring their own business models, which currently mainly include three ways. The first is to provide quantum computing chips or whole machines, and the current buyers are mainly the military and national scientific research institutions. The second is to provide industry quantum solutions, where quantum computing companies work with industry customers to carry out topic research, helping downstream vertical industry customers to provide a complete set of solutions, including quantum algorithms, model optimization, etc. For example, IONQ has signed a contract with UMD National Quantum Laboratory to provide quantum computing services and equipment access, paying $14 million in three years. The third is to provide computing power for potential customers through cloud computing. Most companies with quantum computing hardware have developed cloud platforms, and part of IONQ's revenue comes from QCaaS services, providing users with quantum system solutions and maintenance support.Quantum Computing Applications: Chemical Industry, Finance, and Pharmaceuticals Expected to Take the Lead

The exploration of quantum computing applications is currently focused on three main directions: quantum simulation, quantum combinatorial optimization, and quantum linear algebra. Quantum simulation can simulate the interactions of microsystems at the atomic scale, with application fields including physical models, biopharmaceuticals, and material research; quantum combinatorial optimization uses quantum algorithms to solve polynomial NP-hard problems, suitable for fields such as quantitative finance, traffic planning, and weather forecasting; quantum linear algebra's ability to solve problems involving matrices and vectors is applied in quantum machine learning, cryptography, and more. At present, the application of quantum computing is still in its early stages, but it is widely explored across multiple fields. Based on the level of technological maturity and potential value, the chemical industry, finance, and pharmaceuticals are expected to be the first to benefit.

Specifically, in the chemical industry, quantum computing has broad prospects in chemical synthesis, material design, and energy development, which can help accelerate R&D and reduce costs in the chemical industry; in the financial field, quantum computing will revolutionize portfolio optimization, quantitative trading, simulation pricing, risk prediction, and fraud detection capabilities; in the pharmaceutical field, quantum computing can empower target identification, molecular design, and all stages of clinical trials, improving efficiency and precision. The exploration of practical application scenarios based on NISQ prototypes has become a core research direction in the industry. ICV estimates that from around 2025, quantum computing will begin to release value in industry applications.

Software, Algorithms, and Cloud Services: Solving Real Problems Requires Collaboration Between Hardware and Software

Consistent with classical computers, the quantum computer ecosystem needs to exist around operating systems, compilation software, development tools, and application algorithms and software. To solve real problems, it is necessary to coordinate the development of hardware and software, algorithms, and cloud services. Looking at the development level of quantum computer system software, quantum algorithms, and quantum software technology worldwide, the United States still leads in related technologies and applications in 2022, followed closely by China, Japan, and Canada.

Quantum software is a key component of the quantum computing system, and the current mainstream quantum computers adopt a hybrid system of classical and quantum. Error correction technology is one of the main developments in system software, and hardware manufacturers have all released their own compilation software, with community open-source development becoming the mainstream model in the future. Quantum algorithms can accelerate the solution of certain computational problems and reduce the cost of infrastructure.

Quantum cloud platforms have become a strong promoter of the commercialization process of quantum computing. Due to the immaturity of quantum computer hardware, the large space it occupies, the high manufacturing cost, the strict environmental requirements, the high difficulty of maintenance, and the inability to meet the requirements of individual users for local deployment, quantum computing cloud platforms have gradually become the main form of quantum computing power output in recent years. Quantum cloud services refer to the quantum computing capabilities provided through cloud computing platforms, allowing individual or enterprise users to access quantum computing resources via the internet without owning physical quantum computers themselves.

For the three types of quantum cloud platforms currently on the market, adopters' choices may shift from Q-SaaS to Q-PaaS and Q-IaaS. Q-IaaS provides quantum computing hardware and supporting facilities, allowing users to call hardware resources at a low cost; Q-PaaS provides a development platform where users can develop software on the cloud platform; Q-SaaS provides packaged application service solutions. Currently, most quantum cloud platforms are mainly Q-SaaS, as the technical barriers are low, and enterprises can develop software suitable for themselves through cooperation. With the growth of quantum developers and the improvement of hardware, it is expected that users will gradually shift from Q-SaaS to Q-PaaS and Q-IaaS to meet the needs of data security and differentiated strategies.

Quantum computing cloud platform service providers are developing in a diversified manner, with competition and cooperation coexisting. Currently, cloud platform suppliers are mainly divided into two categories. The first category of suppliers is involved in the research and development of quantum computing hardware themselves, placing their self-developed quantum computers and other hardware on the cloud platform. Typical enterprises include IONQ, PASQAL, Rigetti, and Origin Quantum. The second category of suppliers provides cloud platforms for users by accessing the quantum computing hardware or software systems of other companies. Typical enterprises include Microsoft, Amazon, and Strangeworks. There are also some enterprises that have both capabilities, such as IBM, which launched a commercial quantum computing cloud platform as early as 2016. In addition to being able to call on AQT, IONQ, and other quantum computing hardware resources, it has also developed several superconducting quantum computing chips.Compared to using quantum computers directly, cloud platforms are more cost-effective. Quantum computers are not widely available to the public due to their high cost and manufacturing difficulty. Currently, tech giants such as Microsoft and AWS offer a full range of quantum computing service solutions. Assuming a scientist runs a quantum machine learning algorithm on a Rigetti quantum computer using Amazon Braket, which includes 50 iterations, 2 tasks per iteration, and 100 executions per task, according to our calculations, the minimum cost to run this job is a total of $33.5.