AI has entered a new phase. It is no longer just about building larger models or accessing more data. Today’s competition centers on speed, efficiency and innovation. Companies are seeking new tools that offer both technical and economic advantages. For some, quantum computing is starting to look like one of those tools.

Quantum AI refers to the combination of quantum computing with artificial intelligence. It offers a new way to tackle complex problems in machine learning, optimization and data analysis. While still in development, the potential is drawing serious attention. A 2024 global survey by SAS  found that more than 60 percent of business leaders are already exploring or investing in Quantum AI. However, most also said they do not fully understand what the technology is or how it might be used.

This article explains what Quantum AI is, what problems it might help solve and where it could make an impact in the near future.

Why AI Teams Are Looking at Quantum

Training large AI models takes time, energy and money. Even minor efficiency improvements can result in significant savings. Quantum computing provides new methods for solving certain problems more efficiently or accurately than classical machines.

For example, quantum computers can perform multiple calculations simultaneously, utilizing a property known as superposition. This makes them a good fit for problems that involve searching large spaces or optimizing complex systems. These capabilities align well with many tasks in machine learning, such as feature selection, model tuning and data sampling.

While today’s quantum machines are still evolving, researchers are finding ways to combine them with classical tools. These hybrid systems allow AI teams to test quantum methods now, without waiting for fully developed quantum hardware.

What Quantum AI Is and Isn’t

Quantum AI is not about replacing current AI systems with quantum versions. It is not about running deep learning models entirely on quantum hardware.

Instead, it focuses on using quantum algorithms to support parts of the AI pipeline. These might include tasks like speeding up optimization, improving how features are selected or enhancing sampling from profitability distributions. In these cases, quantum computers do not replace existing tools; they support them.

The work is still experimental. Most examples rely on hybrid methods, in which quantum and classical parts work together. But these systems are already showing results in narrow use cases.

Current Applications Under Development

Although the field is new, Quantum AI is already being tested in several industries. These examples use real tools and published research. They also reflect the kinds of problems quantum methods are best suited to solve.

Model Compression and Feature Mapping

AI models are growing larger and more costly to train. Quantum technologies can help reduce the size and complexity of these models. One method is quantum feature mapping, where input data is transformed using quantum circuits. These transformations can help separate data points that are hard to classify with standard techniques.

In the ‘early’ days a  2021 paper in Nature Physics explored how quantum kernels could improve support vector machines, a type of machine learning model. This approach works well for high-dimensional or sparse datasets, where classical models struggle.

Portfolio Optimization in Finance

Banks and asset managers often use AI to manage portfolios and assess risks. These tasks involve large numbers of variables and constraints. Quantum algorithms like QAOA (Quantum Approximate Optimization Algorithm) are being tested to solve these problems more efficiently.

Citi Innovation Labs and AWS recently studied using quantum computers for portfolio optimization, specifically employing the QAOA algorithm and how it performed. The collaboration showcases the growing interest and investment in quantum computing as a tool for solving real-world problems.

Drug Discovery and Molecular Modeling

Drug development relies on predicting how molecules interact with each other. AI models can assist, but classical simulations have limits. Quantum computing is better suited for modeling chemical systems at the quantum level.

A new study from IBM, The Cleveland Clinic and Michigan State University demonstrated a new way to simulate complex molecules using current-generation quantum computers, offering a viable path forward for quantum-centric scientific computing.

Supply Chain Optimization

Supply chains are difficult to manage due to their size and complexity. AI can help, but certain tasks, like route planning and inventory control, remain hard to optimize. Quantum methods are being explored to improve these tasks.

Fujitsu partnered with Japan Post to optimize last-mile delivery in Tokyo, where traditional routing algorithms failed to account for dynamic variables like traffic congestion and parcel volume fluctuations. By deploying Quantum AI, they were able to start work on transforming some of the most fundamental aspects of logistics.

Challenge and Limitations

Quantum hardware remains a challenge. Although there are new advancements seemingly every day, today’s machines are still sensitive to noise, hard to scale and unreliable for long computations. Most applications must operate within these limits, utilizing shorter and simpler quantum circuits.

Quantum software development is also difficult. Quantum programming requires knowledge in physics, mathematics and computer science. Few teams have the right mix of skills.

To lower this barrier, new tools are being created. These include high-level programming frameworks and automated circuit design systems. These allow AI developers to test quantum methods without needing to write low-level quantum code.

What AI Teams Can Do Today

Quantum AI is not ready for full deployment. However, forward-looking teams can begin building the knowledge and systems needed to take advantage of it in the future. Here are three steps to consider:

Build cross-functional teams – Combine AI experts with researchers in optimization and quantum computing. This allows teams to explore new ideas and prepare future capabilities.
Experiment with hybrid workflows – Focus on narrow problems where quantum components can support classical models. These include feature selection, sampling or constrained optimization.
Use tools that abstract complexity – Adopt platforms and frameworks that hide low-level quantum details. These tools help teams focus on the application, not the hardware.

Quantum AI is still developing. It is not a shortcut or replacement for classical AI. However, it is a growing field with real potential in areas where current models fall short or struggle. The most likely path forward is not sudden disruption, but steady integration.

As quantum hardware improves and software becomes more accessible, early adopters will be better positioned to utilize these new tools. For teams already working at the limits of classical systems, Quantum AI may be the next place to find value.