For decades, quantum computing has occupied an unusual place in technology. Its theoretical potential has been enormous, while the machines themselves have remained difficult to build, error-prone, and far removed from the reliability of conventional computers.

That gap is beginning to narrow.

Researchers have made important advances in quantum error correction, companies are improving the hardware and software needed to operate quantum systems, and businesses in fields such as finance, pharmaceuticals, energy, and manufacturing are experimenting with quantum applications. At the same time, governments and technology providers are preparing for a future in which powerful quantum computers could undermine widely used forms of encryption.

None of this means that quantum computers are about to replace conventional computing. They are more likely to emerge as specialized processors used alongside CPUs, GPUs, and supercomputers, tackling particular problems for which quantum mechanics provides a computational advantage.

That distinction is important. Quantum computing is becoming more commercially relevant, but much of its economic potential still depends on advances that have not yet been achieved.

Traditional computers store and process information using bits that take a value of either 0 or 1. Quantum computers use quantum bits, known as qubits, which can be prepared in combinations of quantum states known as superpositions.

Quantum algorithms also make use of phenomena including entanglement and interference. By carefully controlling these effects, a quantum computer can manipulate information in ways that are fundamentally different from conventional computation.

This does not mean a quantum computer simply tries every possible answer simultaneously. Quantum algorithms must be specifically designed so that interference increases the probability of useful outcomes while suppressing others.

The result is a machine that could eventually outperform classical computers on certain types of problems, while offering little or no advantage on many everyday computing tasks. Email, spreadsheets, websites, most databases, and conventional business software have no obvious reason to move to quantum machines.

The opportunity instead lies in computational problems where the structure of the problem matches what quantum algorithms can do well.

Qubits are extremely sensitive to their environment. Small disturbances can introduce errors, making long and complicated calculations difficult to execute reliably.

Quantum error correction attempts to solve this problem by encoding information across groups of physical qubits to create more reliable logical qubits. The objective is not simply to build machines with more qubits, but to reduce errors enough that useful calculations can continue for the enormous number of operations some quantum algorithms require.

This is why headline qubit counts should be interpreted cautiously. A physical qubit, a logical qubit, and a qubit built using a different hardware architecture are not directly interchangeable measures of computing capability. Error rates, connectivity, operating speed, control quality, and error-correction overhead all matter.

One important milestone came from Google Quantum AI's Willow processor. Research published in Nature showed that increasing the size of its surface-code error-correction system reduced the logical error rate rather than making it worse. The experiment included a distance-7 logical memory built with 101 physical qubits. The researchers emphasized that the result demonstrated an important error-correction threshold, not a finished fault-tolerant computer.

Progress continued in 2026. Research published in Nature in July showed a system that used reinforcement learning to continually adjust a Willow processor while error correction was running. The method improved the logical stability of the tested surface code by 3.5 times under deliberately introduced drift.

In October 2026, researchers reported another step forward in Science. A technique known as magic state cultivation reduced errors in specially prepared quantum states by a factor of 40 and achieved a measured state fidelity of at least 99.99 percent in the experiment. Magic states are needed for important operations in many fault-tolerant quantum computing architectures, although the researchers still need to demonstrate the technique at much larger scale.

Together, these developments show why the industry has become more optimistic. They also illustrate how much engineering remains between a research breakthrough and a machine capable of operating reliably on large commercial problems.

Researchers have already demonstrated quantum computers outperforming classical machines on carefully designed computational tasks. The harder challenge is turning that performance into something economically useful.

Google's Quantum Echoes experiment, announced in 2025, provided an important example. The company reported that its Willow processor executed a verifiable quantum algorithm approximately 13,000 times faster than the classical algorithm used for comparison. Unlike earlier benchmark demonstrations, the method also has potential connections to the study of molecular structures. However, the molecular simulations demonstrated alongside the work were still small enough to be handled classically, so the experiment did not establish an end-to-end commercial quantum advantage.

That distinction reflects the broader state of the industry.

A September 2026 commentary in Nature Electronics described the central challenge as moving from engineered demonstrations of quantum advantage toward calculations with genuine practical value.

The technology is therefore crossing important scientific thresholds before it has crossed the economic one.

The most realistic path toward commercial quantum computing does not involve replacing today's data centers with quantum machines.

Instead, quantum processors are increasingly being developed as part of hybrid computing systems.

A conventional computer might handle data processing, storage, optimization, and software control. GPUs or high-performance computing systems could perform large numerical workloads. A quantum processor would then be called upon for a specific part of a calculation where quantum methods appear advantageous.

This model is already reflected in commercial cloud services. Amazon Braket, for example, provides access to several types of quantum hardware, including trapped-ion, superconducting, and neutral-atom systems, along with classical simulators and tools for running hybrid quantum-classical algorithms. Customers can experiment with the technology without purchasing a quantum computer themselves.

Research applications are following the same approach. In 2026, researchers from Cleveland Clinic, RIKEN, and IBM reported a heterogeneous quantum-classical chemistry calculation involving biologically relevant systems and a largest modeled system containing more than 12,000 atoms. Quantum processors worked alongside classical supercomputers rather than replacing them. The result was a research demonstration, not proof that quantum computing had commercially outperformed the best classical method for the entire problem.

For businesses, this could make quantum adoption look more like adding a specialized accelerator to the computing stack than replacing existing infrastructure.

One of the strongest arguments for quantum computing comes from physics itself.

Molecules and materials are quantum systems. As their complexity increases, accurately describing their electronic behavior can become exceptionally demanding for classical computers. A sufficiently capable quantum computer could potentially represent some of these systems more naturally.

That possibility has attracted interest from pharmaceutical companies, chemical producers, battery developers, manufacturers, and materials researchers.

Potential applications include modeling molecular interactions, investigating reaction pathways, designing catalysts, studying battery materials, and screening potential drug compounds.

The commercial impact remains uncertain because classical computational chemistry is also improving rapidly. Research published in Nature Communications has cautioned that evidence for a generic exponential quantum advantage across quantum chemistry has not yet been established. Quantum computers could still deliver valuable improvements, but the size and form of the advantage will depend on the particular problem and algorithm.

This is an important limitation. The business case for quantum chemistry should not be based on the assumption that every molecular simulation will suddenly become dramatically faster.

Many industries face optimization problems in which enormous numbers of possible combinations must be evaluated.

Airlines schedule aircraft and crews. Logistics companies plan routes. Manufacturers coordinate production. Utilities balance electricity networks. Financial institutions construct portfolios under multiple constraints.

Quantum algorithms are being investigated for these problems because some may be able to explore complicated mathematical structures more efficiently than existing techniques.

But optimization is also one of the areas where claims of quantum advantage require particular caution. Classical optimization software is highly sophisticated, and specialized algorithms often perform extremely well. A quantum solution must therefore outperform not merely a theoretical classical approach, but the best practical alternatives available at the time.

For this reason, current work frequently concentrates on hybrid methods in which quantum processors handle selected subproblems while classical systems perform the rest of the calculation. McKinsey's 2026 Quantum Technology Monitor identified transport and logistics among the areas where companies are testing this type of hybrid approach.

Financial institutions have also emerged as prominent quantum computing researchers.

Potential applications include portfolio construction, risk analysis, optimization, machine learning, and cryptography. JPMorganChase, for example, maintains researchers working on quantum algorithms related to optimization, artificial intelligence, cryptography, and finance.

In 2025, researchers from JPMorganChase, Quantinuum, U.S. national laboratories, and the University of Texas at Austin demonstrated a protocol for generating certified quantum randomness. Randomness is important in cryptography and other computational processes, and the research showed a task that exploited properties of quantum systems rather than simply using a quantum computer as a faster conventional processor.

Financial applications such as portfolio optimization remain largely experimental. The significance lies less in current deployment at scale than in the willingness of major institutions to develop expertise before more capable hardware becomes available.

Quantum computing does not have to become commercially useful for molecular simulation or optimization before it affects corporate technology strategies.

Cybersecurity is already changing.

Large enough fault-tolerant quantum computers could run algorithms capable of threatening widely used public-key cryptography. The machines required to do this do not exist today, and the date at which they might arrive remains uncertain.

The migration away from vulnerable cryptography, however, can take years.

In August 2024, the U.S. National Institute of Standards and Technology finalized its first three principal post-quantum cryptography standards. They include algorithms for establishing encryption keys and creating digital signatures that are designed to remain secure against future quantum attacks. NIST now says organizations should begin migrating to quantum-resistant cryptography.

NIST's transition planning calls for quantum-vulnerable public-key algorithms to be removed from its standards by 2035, with higher-risk systems expected to move earlier.

The urgency also reflects the possibility of what security researchers call harvest now, decrypt later attacks. An adversary could collect encrypted information today and retain it until a future computer becomes capable of breaking the encryption. Data that must remain confidential for many years therefore faces a potential quantum risk well before such a computer actually exists.

For many companies, preparing cryptographic systems for the post-quantum era may become the first tangible business consequence of quantum computing.

Investment and corporate experimentation are expanding even though large-scale fault-tolerant quantum computing has not yet arrived.

McKinsey's 2026 Quantum Technology Monitor identified more than 300 organizations worldwide working with quantum computing companies and analyzed 162 of them in greater detail to examine adoption and use cases.

The report estimates that quantum computing companies generated more than $1 billion in worldwide revenue in 2025. McKinsey projects that this could rise to as much as $4.4 billion by 2028. These figures are market estimates rather than audited industry sales totals. McKinsey says its analysis combines public information, expert interviews, proprietary research, market databases, investor reports, and company announcements.

Its longer-term estimates are much larger but measure something different.

McKinsey projects an internal quantum computing market, covering hardware, software, and services, of $43 billion to $71 billion by 2035. Separately, it estimates that quantum computing could generate $1.3 trillion to $2.7 trillion in economic value across industries by the same year. Economic value refers to potential benefits created through quantum-enabled applications, not revenue earned by quantum computing companies.

The wide ranges reflect substantial uncertainty about technological progress, adoption, and the applications that ultimately prove commercially useful.

Company roadmaps have become increasingly ambitious, but they should not be confused with delivered capabilities.

IBM's current roadmap, for example, targets a fault-tolerant system called Starling for 2029, with 200 logical qubits and the ability to execute 100 million quantum gates. IBM describes these figures as future objectives that are subject to change.

The U.S. Defense Advanced Research Projects Agency is taking a broader approach. Its Quantum Benchmarking Initiative is evaluating competing technologies to determine whether an industrially useful quantum computer can be built by 2033. DARPA defines utility scale as a system whose computational value exceeds its cost.

As of October 2026, Atom Computing, Diraq, IBM, IonQ, Microsoft, and PsiQuantum had reached the initiative's final evaluation stage, using approaches ranging from neutral atoms and trapped ions to superconducting systems, silicon spins, topological designs, and photonics. DARPA explicitly states that the program is not intended to select a single winner and that it remains uncertain which architecture, if any, will ultimately reach utility scale.

That diversity is another reason raw qubit numbers can be misleading. The industry has not yet settled on one dominant technology comparable to the silicon transistor architecture that eventually standardized conventional computing.

For most companies, preparing for quantum computing does not require buying quantum hardware.

The more immediate task is understanding where the technology could intersect with existing operations.

Companies with long-lived sensitive information can begin identifying cryptographic systems that will eventually require migration to post-quantum standards. Businesses in chemistry, pharmaceuticals, finance, logistics, manufacturing, and energy can identify computational problems that are expensive or difficult today and determine whether credible quantum algorithms exist for them.

Cloud services also make experimentation possible without large capital commitments.

The objective should not be to force quantum computing into problems that classical technology already solves efficiently. It should be to identify areas where better computation would create measurable economic value and then track whether quantum hardware and algorithms become capable of delivering it.

Quantum computing is unlikely to transform technology by making existing computers obsolete.

Its more plausible impact is as a new layer of specialized computing.

CPUs will continue to run general software. GPUs and other accelerators will remain central to artificial intelligence and high-performance computing. Quantum processors could eventually join that ecosystem, performing selected calculations that conventional machines struggle to handle efficiently.

The scientific progress behind that possibility is becoming increasingly concrete. Error correction is improving, specialized quantum advantages have been demonstrated, cloud access is expanding, and companies are building applications around increasingly capable hardware.

Yet the most important commercial threshold remains ahead. Researchers still have to turn quantum advantage on carefully selected experiments into reliable, repeatable economic advantage on problems that matter outside the laboratory.

If they succeed, quantum computing may not replace the computers businesses use today. It could instead expand the boundaries of what those computers are capable of solving.