Quantum Computing vs. AI: Real-World Applications

While artificial intelligence dominates the current wave of innovation in finance, another technology is beginning to attract attention: quantum computing. Financial institutions are already testing it for portfolio optimization and trade-execution prediction and exploring its potential for Monte Carlo option pricing.

Up until this point, every computer whether used in a trading desk terminal or running large language models (LLMs) processes information in the same “classical” way: as a sequence of binary bits, each resolving into 0 or 1. That architecture has powered decades of financial analysis. Even the latest agentic AI systems ultimately rely on classical hardware.

But this also imposes a ceiling: certain problems (e.g., optimizing hundreds of securities in a portfolio under real-world constraints, modeling correlated tail risks across asset classes, and pricing derivatives whose complexity resists closed-form solutions) grow exponentially harder as they scale, quickly outpacing what classical supercomputers can handle. Quantum computers work via fundamentally different physical principles that could lift that ceiling dramatically.

Researchers describe the potential breakthrough of quantum computing as quantum advantage: the point at which quantum computers can perform a computation “more accurately, cheaply, or efficiently than a classical computer” (IBM, 2025). In investment management, where even small improvements in forecasting, optimization, or risk modeling can translate into millions of dollars, even modest computational gains matter. But quantum computing raises the possibility of improvements that are not incremental or linear but qualitatively different, in some cases scaling exponentially.

Early demonstrations hint at what such advantages could look like in practice. Recent experiments have shown quantum systems performing certain high-dimensional calculations that would be impractical to reproduce with classical simulations.

In October 2025, Google announced that its 105-qubit “Willow” processor completed a task-specific computation roughly 13,000 times faster than the best available classical supercomputer simulations. The task was completed in just over two hours, compared with an estimated 3.2 years on one of the world’s most powerful classical supercomputers.

Quantum advantage therefore represents more than a new technology following AI. It signals the possibility of an altogether different computational paradigm. It opens the door to solving problems that no conventional computer could ever fully resolve, regardless of how much processing power you threw at them, including some of the most stubborn computational challenges in investment management.

This new paradigm is especially consequential in a future world of AI agents. Quantum computers could not only accelerate certain AI computations, but also introduce fundamentally new ways of representing uncertainty, correlations, and high-dimensional relationships. In doing so, they expand the kinds of problems AI systems can meaningfully learn from and analyze.

That said, quantum technologies, and quantum computation in particular, are unlikely to become as ubiquitous as classical computers. Rather than something in everyone’s hands, quantum systems will likely be applied to highly complex and computationally demanding problems. For institutional investors and asset managers, that includes challenges such as scenario modeling or portfolio optimization under many constraints.

While everyday investors may not be leveraging quantum anytime soon, banks, hedge funds, asset managers, and pensions have already taken notice of the potential of quantum computing.

In late 2025, Vanguard announced a partnership with IBM “to explore how quantum computing can revolutionize portfolio construction, one of the most complex challenges in financial management.” They will do so by applying “hybrid quantum-classical algorithms to simulate dynamic markets and optimize portfolios under real-world constraints like liquidity, transaction costs, and regulatory limits.”

Around the same time, HSBC announced empirical evidence of real-world advantage of using quantum computing for algorithmic bond trading. Also partnering with IBM, HSBC found a 34% improvement over current computer models in predicting the likelihood of a trade to be filled at a quoted price.

Among asset owners, Canadian pension fund BCI recently partnered with Quantum Algorithms Institute, a British Columbia-based non-profit, “to identify quantum investment applications for portfolio optimization, risk assessment, and financial modeling, while implementing post-quantum security standards to support BCI’s long-term operational resilience.”

Alongside these partnerships, the quantum computing ecosystem is maturing rapidly. Cloud-accessible quantum computing platforms such as IBM Quantum, Azure Quantum (Microsoft), and Amazon Braket have opened quantum experimentation to financial data scientists. Across North America and Europe, national initiatives are investing billions to secure leadership in quantum computing, quantum sensing, and quantum communication, including cryptographic methods.

A recent MIT report shows that venture funding for quantum technologies increased drastically since 2021. The momentum across hardware, software, enabling technologies, and regulatory policy underscores that the era of practical quantum experimentation has already begun.

But given that we have not yet reached quantum advantage and commercial-level viability, many still wonder whether quantum computing is an actual emerging technology or more of a theoretical fantasy. We aim to address these concerns.

Quantum computers are systems that use qubits as opposed to classical bits. Instead of using 0 and 1 as classical computers do, qubits exploit quantum phenomena such as superposition¹ and entanglement² to perform computations. In doing so, qubits have the potential to drastically increase the available compute beyond what is possible using classical bits.

Current quantum computers have demonstrated — both theoretically and experimentally — that they offer capabilities beyond classical computation for certain types of problems. BBVA, Goldman Sachs³ and JPMorgan Chase have already developed prototype quantum algorithms that could be applied in finance, whether for portfolio optimization, data security, or solving other complex business problems. However, large-scale, fault-tolerant quantum computers remain out of reach due to challenges of noise and error-correction.

Yet, promising results have come from hybrid quantum computing, in which part of a computational task runs on a quantum processor while the remainder runs on classical hardware. Thanks to this hybrid approach, the estimated timeline for achieving quantum advantage for tasks like portfolio optimization and Monte Carlo option pricing have decreased from a decade to fewer than five years.

Consider the transition from CPUs (central processing units) to GPUs (graphics processing units) that has become synonymous with greater computational power and the advancement of AI over the past decade. Quantum computing has its own equivalent of processing units: QPUs (quantum processing units). Their architectures vary widely. There are multiple methods for developing QPUs, and they all rely on the physical properties of the qubits that serve as the fundamental units of quantum information.

Several prominent QPU architectures⁴ include:

There remains no decisively “best” approach among these methods, and companies generally commit resources to the approach they believe to be most promising.⁵ Additionally, each of these architectures implies different timelines for commercial maturity, influencing when quantum applications will reach widespread enterprise use.

Beyond competition among companies, however, countries are also racing to achieve high-performance, commercially available QPUs. According to MIT’s Quantum Index Report 2025, there are approximately 40 commercially available QPUs around the world with more than 80 companies currently developing quantum hardware.

But hardware is just one consideration. Software stacks that connect quantum hardware to programming tools are also evolving rapidly. These systems manage low-level operations and enable developers to build and run algorithms tailored to business and scientific applications.

Most software stacks are cloud-accessible on everyday classical computers. IBM’s Qiskit, for example, is an open-source software development kit (SDK), allowing finance R&D teams to write, simulate, test, and — via IBM Quantum services — run quantum algorithms on actual quantum hardware through web-based and cloud interfaces. Several companies offer hybrid services through API calls or other methods rather than SDKs. Quantum algorithms can be run remotely on quantum hardware through classical systems.

Popular quantum SDKs include:

• Qiskit (IBM)
• Cirq (Google)
• Braket SDK (Amazon)
• Azure Quantum SDK (Microsoft)
• PennyLane (Xanadu)

On top of this software stack are platforms that offer SaaS (software-as-a-service) turnkey quantum algorithms and workflows. QC Ware Forge, for example, is built on Amazon Braket and connects to various QPUs and simulators. QC Ware Forge can run a range of optimization and machine learning tasks, such as classification, that can be useful for investment management and risk management. Other companies such as Multiverse Computing have developed similar offerings, integrating with quantum hardware providers to offer ready-to-use quantum algorithms targeted to financial optimization problems (among others).

Quantum computing is an emerging technology competing with a multi-trillion-dollar classical computing industry backed by more than eight decades of continuous progress. It would be unrealistic to expect quantum computers to outperform classical machines across all domains. Current quantum hardware still faces substantial limitations, and this is widely acknowledged. Here, we focus on the most evident physical challenge: the fragile nature of quantum systems.

Qubits and the quantum gates operating on them must be isolated to prevent interaction with undesirable external factors such as radiation, temperature fluctuations, and system vibrations. All these uncontrolled interactions constitute what is known as noise.

Due to these technical challenges, the number of qubits and gates available for practical applications is limited. And many qubits must be devoted to compensating for noise-induced errors, a process known as error correction. To address these hardware constraints, researchers are developing more sophisticated error correction methods.

These physical constraints affect all major hardware platforms under development, whether based on superconducting qubits, photonics, neutral atoms, or any other architecture mentioned above. As quantum processors scale up, they inevitably face trade-offs involving qubit quality and gate performance.

Far from being insurmountable, these challenges have inspired quantum experts to develop innovative algorithms tailored for currently noisy intermediate-scale quantum (NISQ) devices. This new breed of systems is being actively employed to solve financial problems such as risk management, option pricing, and a variety of computationally expensive tasks specific to finance.

A common misconception is that fault-tolerant quantum computers — those capable of correcting errors in real-time and providing reliable computations, analogous to current classical computers — will need to fully replace classical computers to deliver significant computational advantages. In reality, hybrid quantum-classical computing may offer near-term advantages without requiring such technological maturity.

One important hybrid system use case is for portfolio optimization. D-Wave recently published results showing that their hybrid quantum processor combined with classical solvers could outperform classical algorithms (including those used in machine learning) in multi-asset portfolio optimization tasks that include large, indivisible assets like private equity and real estate. Outperformance was obtained by backtesting for several risk-return measurements, including accumulated return, annualized return, Sharpe ratio, and the Sortino ratio across multiple time horizons.⁶

But hybrid systems are not the only option researchers are exploring amid the challenges of scaling quantum hardware. Financial quantum computing teams are increasingly developing quantum-inspired methods — algorithms that run entirely on classical machines but borrow key concepts from quantum computation.

These methods selectively incorporate ideas from quantum mechanics to reshape how classical algorithms search, optimize, or compress information. As a result, quantum-inspired approaches can sometimes outperform conventional classical algorithms on specific optimization or linear-algebraic problems, even though they run on standard processors or specialized accelerators. Their advantages are narrower than what fully quantum algorithms are expected to achieve once fault-tolerant hardware scales. But they are practical today because unlike even hybrid systems, they require no quantum hardware and can be applied at full industry scale.

A similar strategy has emerged in the field of post-quantum cybersecurity. Technologies such as quantum key distribution (QKD) — a purely quantum solution for secure communication against future quantum threats — still face significant engineering hurdles that limit their scalability. For that reason, cybersecurity experts have developed classical cryptographic protocols designed to withstand attacks from future quantum computers. 

These post-quantum cryptography standards are being deployed worldwide, demonstrating that effective quantum-aware solutions can be implemented long before universal quantum computers exist.

This discussion leads to a practical question for finance professionals: What should their next steps be, and how can they move from passive awareness to active exploration of quantum solutions within their organizations?

Finance professionals can take several steps to begin developing their quantum capabilities. In each case, the challenge is not to get lost in the technical details, but rather, observe the bigger picture: where the technology is currently applied, promising avenues for development, and how quantum computing is likely to change the industry. Ultimately, the coming quantum era in finance will be an evolution, not an overnight revolution, so there’s no need to panic.

However, to prepare adequately for the evolution, interested financial professionals should:

• Build focused quantum literacy
• Map the technology to real business problems
• Experiment with hybrid and quantum-inspired approaches
• Get ahead of post-quantum cybersecurity

Footnotes

^[1] Superposition refers to how qubits generate multidimensional computational spaces utilizing features of quantum mechanics that allow probabilistic distributions across all possible configurations (see IBM).

^[2] “Entanglement is the ability of qubits to correlate their state with other qubits” (IBM).

^[3] To see the details of Goldman Sachs and Quantum Motion research on quantum computing for option pricing, see https://journals.aps.org/pra/pdf/10.1103/m32k-7nq2.

^[4] Detailed descriptions of quantum computing modalities and architectures can be found on PostQuantum by Martin Ivezic.

^[5] For a list of current quantum computers, the type of method used, and the number of qubits, see Augmented Qubit by Kihara Kimachia (as of this post, last updated July 2025).

^[6] See also research from Vanguard and IBM that demonstrates using a quantum-classical workflow can achieve better accuracy than purely classical approaches when optimizing construction of an ETF.