Quantum computing unlocks a completely new way of computing using the principles of quantum physics. As quantum computers mature, they will be integrated with supercomputers to solve some of the world's most challenging computational problems.

Today, industries such as automotive, pharmaceutical, chemical, and financial services are exploring how to use quantum computing as an accelerator for traditional supercomputing.

What is Quantum Computing?

Quantum computing is a paradigm-shifting technology that leverages the laws of quantum physics to tackle extremely difficult problems. These problems are so computationally intensive that solving them on traditional supercomputers requires vast resources.

At the heart of quantum computing is the quantum bit, or qubit. While classical bits exist only in states 0 or 1, qubits can exist in a so-called superposition of both states. This means that N qubits in superposition hold information related to an exponential number (2N) of binary configurations, which together constitute a quantum state. When an operation is performed on any one of the N qubits, the entire quantum state is manipulated—indicating enormous parallelism. However, applying this capability is subtle, as reading information from the quantum state can only be achieved by probabilistically measuring individual configurations after the computation.

To fully utilize quantum parallelism, quantum applications must additionally exploit the properties of entanglement and interference.

Quantum Computing Applications

Although only certain applications have confirmed quantum algorithms that can effectively solve significant problems, these algorithms do offer solutions that cannot be generated by traditional supercomputing alone.

Therefore, quantum computing is expected to impact numerous industries, including defense, energy, logistics, engineering, medicine, finance, and retail.

Some demonstrated examples include:

• Simulating complex chemical systems to guide the discovery of new batteries, solar cells, pharmaceuticals, and consumer products.
• Faster estimation of financial forecasts and risk metrics to help improve trading margins and strengthen risk management.
• Optimizing complex systems, including global supply chains and genetics.
• Providing offensive and defensive cryptographic capabilities, changing the landscape of cyber defense and national security.
• Developing distinct AI programs to extract deeper insights from data and provide more reliable predictions for decision-makers.

These applications have varying requirements for quantum computing. Therefore, some industries are expected to be early adopters of quantum computing applications. The fields of chemistry and materials applications are expected to benefit first from first-generation quantum devices, while other fields will require larger-scale, next-generation quantum computers.

Significant efforts are being made by governments, academia, and industry to identify the most impactful use cases and to build new quantum algorithms to solve such problems. The goal of quantum application researchers remains to discover more applications that can demonstrate "quantum advantage." The most valuable applications of quantum computing are yet to be discovered.

What is a Quantum Computer?

A quantum computer is a device capable of isolating and manipulating qubits. Qubits are physical objects with controllable quantum properties. Achieving quantum computation requires advanced engineering and state-of-the-art supercomputing-assisted infrastructure to manipulate qubits. A Quantum Processing Unit (QPU) is a device capable of carefully interacting with multiple qubits, often through methods such as electrical signals, microwaves, radio frequency (RF), or laser pulses.

There are various candidate physical objects that can be used to build qubits, such as superconducting coils, neutral atoms, trapped ions, electron spins, nitrogen-vacancy centers in diamond, photons, or other specialized materials. Each qubit has its own advantages and disadvantages. For example, some qubits need to be cooled to near absolute zero, thus requiring specialized cryogenic devices such as dilution refrigerators. Other types of qubits require ultra-high vacuum environments to prevent the fragile states of the qubits from being disturbed by environmental noise.

Qubits need to be effectively isolated from the environment to ensure their quantum properties remain intact and are not destroyed by the slightest environmental disturbance. This can be achieved through conventional electrical hardware (i.e., control systems), which must be able to manipulate the qubits and read their information during algorithm execution. A key component of this control system is a conventional supercomputer, used to coordinate control operations, correct errors, and analyze the output of the QPU. Here, some applications are inherently hybrid, meaning they require the combined use of conventional computers and QPUs to complete higher-level workflows.

A key challenge faced by hardware developers is designing a quantum computing system whose operating components can be scaled to handle larger computational tasks while ensuring that the qubits remain protected and controllable.

Quantum Computing Programming

Quantum programs are typically written in common languages such as Python or C++ using quantum development frameworks. Most frameworks allow users to specify basic quantum operations (gates). Some frameworks take a more functional approach and also allow users to specify high-level kernels (sometimes called "oracles"). Such kernels encapsulate the large number of gates required to perform more advanced functions, meaning that programmers do not have to worry about gate-level details. Visually, programs are often represented by quantum circuits composed of gates or commonly used kernels.

To run a program on a quantum computer, it must be compiled to translate the abstract gate-level operations into a set of device-specific instructions. This compilation task is performed using a conventional computer. Different quantum devices have unique compilation requirements and instruction sets; therefore, an excellent development platform for writing quantum programs should be able to compile the same code into various instruction sets on quantum devices as much as possible, including device-specific optimizations.

The compiled code will be executed on the device's quantum hardware by its control system. In future large-scale quantum computers, this execution phase will also include additional control instructions for implementing quantum error correction (see the next section). With quantum error correction, large quantum programs can be completed without noise interference.

At the end of a quantum program, the qubits in the QPU are read out. In some algorithms, this readout operation is performed probabilistically, sometimes requiring the algorithm to be repeated multiple times. In most algorithms, the readout results require post-processing by a conventional computer to interpret the data as a meaningful solution. Finding quantum circuits that can generate meaningful results using as few qubits and gates as possible is one of the main challenges in quantum algorithm design.

Quantum Error Correction

Currently, noise that makes it difficult to perform quantum computations is mainly eliminated by two methods. One method uses Noisy Intermediate-Scale Quantum (NISQ) devices, which do not attempt to correct noise but instead try to operate within the strict limitations it imposes. The other method uses Fault-Tolerant Quantum Computers (FTQC), which actively implement quantum error correction protocols to try to prevent noise. Because noise accumulates quickly, NISQ devices cannot run algorithms with many qubits or gate operations. Whether any useful applications can run under the constraints of NISQ devices is still uncertain, but researchers continue to explore such algorithms.

While technically more difficult to implement, FTQC is the "holy grail" of the entire quantum ecosystem. Unlike NISQ, some useful applications have been developed for FTQC devices, leading to the general belief that fault tolerance is a practical requirement for any quantum computer. FTQC uses Quantum Error Correction (QEC) to circumvent noise. QEC uses error-correcting codes to map many noisy physical qubits to a single noise-free logical qubit. Similarly, logical operations can be designed to act on these encoded logical qubits. A key element of quantum error-correcting codes is the repeated measurement of qubits, which generates information so that a conventional computer can "decode" where errors have occurred and then interpret those errors.

Implementing QEC is challenging. First, for QEC to work, the noise in the quantum hardware needs to be below a certain error rate "threshold." Second, QEC codes require many physical qubits to encode a single logical qubit (the exact number depends on the hardware and the specific noise constraints of the code used). Finally, QEC codes need to be run repeatedly within "code cycles." The data continuously generated by QEC codes must be processed by a conventional computer under strict time constraints to prevent data backlog from affecting performance. This requires the use of state-of-the-art supercomputers closely coupled with the QPU. Researchers are still determining the QEC codes and optimal fault-tolerant algorithms that minimize hardware overhead.

Quantum Supercomputing

In the popular imagination, the physical size of a quantum computer is comparable to a supercomputer center. The quantum computer's auxiliary components for shielding qubits, cooling qubits, and sending control pulses are large and require a large size to operate precisely on such small systems. Furthermore, quantum computers are designed to solve specific types of problems and cannot efficiently handle many of the tasks performed on a laptop. This means that quantum computers will become accelerators for supercomputers when solving large-scale problems and will naturally become permanent fixtures in data centers.

Quantum computers not only help supercomputers solve unprecedented problems but also rely on supercomputers to operate. This interrelationship is called quantum-accelerated supercomputing. Because quantum computers require supercomputers to perform compilation, control programs, error correction, calibration, and many other tasks, they will always be closely coupled with supercomputers. For some tasks, low latency is critical, so the location of quantum and conventional processors needs to be very close. Supercomputing centers around the world have begun to integrate QPUs and are being used to research various aspects of hybrid quantum-classical computing.

Quantum Computer Simulation

Simulation is an invaluable design tool for building any complex system. For every chip manufactured by NVIDIA, simulation is a key part of the entire development process, helping to shorten development cycles and improve cost-effectiveness. This is even more true for quantum computing hardware, which involves cutting-edge components that can only be understood through a deep understanding of their underlying physical mechanisms. Many quantum components also utilize special or underdeveloped materials, and because it is extremely difficult or expensive to conduct experiments with them, simulation is essential.

Simulation is also crucial for developing and testing quantum algorithms, helping to evaluate algorithm performance. This is especially important because the supply of quantum computers is limited, noisy, access is restricted, and they are expensive, meaning that testing on physical hardware is both impractical and infeasible. Simulation can mimic noise in real devices or provide ideal noise-free results. This theoretical data provides an important benchmark for evaluating actual quantum computing output, which can help understand noise patterns and even offer suggestions for noise mitigation.

Importantly, powerful simulation tools are easily accessible to developers and scientists, both novice and experienced, through learning practical quantum programming tutorials.

AI and Quantum Computing

Quantum computing and AI are poised for a collaborative relationship. On the one hand, quantum computing can accelerate AI applications and even potentially generate entirely new applications. Many such applications are broadly classified as quantum machine learning (QML). QML is an emerging field; currently, the known quantum algorithms that can improve traditional AI require hardware several generations beyond the first useful quantum computers.

However, leveraging AI to improve quantum computing (often called quantum AI) is expected to become a reality in the near future—and is already being utilized. AI is being used to help develop and operate quantum computers. New quantum algorithms and hardware performance are developed using AI technologies, which in turn accelerates the quantum computing roadmap. AI can also be used to address the challenges of operating quantum computers, including achieving more efficient error correction, calibration, device control, task scheduling, and circuit compilation.

AI is becoming a crucial pillar of quantum computing and is a major motivation for the recent integration of quantum computing hardware into some of the world's largest supercomputing centers.

Do Quantum Computers Exist?

Existing quantum computing devices are capable of preparing, manipulating, and reading out hundreds of qubits, and this number will soon reach thousands. Although these devices employ very cutting-edge engineering, they are still in the experimental and demonstration phase. More specifically, current quantum computers cannot compete with traditional computers in performing meaningful, useful tasks.

In a narrow sense, quantum advantage usually refers to the performance of a quantum computer exceeding that of a traditional computer. However, the U.S. government supports the use of a more appropriate term, "practical-scale quantum computing"—emphasizing that as a solution, quantum computers must be cost-effective to be useful. In this context, "cost" is a broad concept and can refer to any meaningful economic, scientific, social, or environmental considerations.

Current devices are still unable to transform entire industries, mainly because their hundreds (or increasingly thousands) of qubits are negatively affected by noise. To achieve practical-scale quantum computing, the development of quantum hardware, error correction techniques, and algorithms must continue to progress. At the same time, students, industry professionals, and field scientists must be quantum literate. Now is the best time to develop educational content that can be used to practice quantum programming skills and more fully understand what problems useful quantum computers are suited to solve.

*Copyright of images or videos (complete or partial) related to NVIDIA products belongs to NVIDIA Corporation.